Skip to main content

Full text of "Journal of the Society of Motion Picture Engineers"

See other formats


From the collection of the 



n 



m 



o Prejinger h 
u v e JJibrary 
t p 



San Francisco, California 
2007 



UME XXVIII NUMBER ONE 

JOURNAL 

of the 

SOCIETY OF MOTION 
PICTURE ENGINEERS 






JANUARY, 1937 



SHED MONTHLY BY THE SOCIETY OF MOTION PICTURE ENGINEER 



SPRING CONVENTION 

Society of Motion Picture Engineers 

Hotel Roosevelt 

Hollywood, Calif. 

May 24th to 27th, Inclusive 



Technical Sessions 

In the New Supper Room, schedule to be announced later. Several sessions wUl 
be held in the evenings to permit those to attend who might otherwise be unable 
o do so Symposiums, General Discussions, Lectures, Demonstrations, and 
Open Forums, dealing with all phases of motion picture techmc. 

Semi-Annual Banquet 

Wednesday evening, in the New Supper Room-Dining, Dancing, Music, 
Floor Shows. Addresses by prominent speakers. 

Hotel Rates 

Minimum rates and excellent accommodations guaranteed to members 
Single $3.50 per day; double, $5.00 per day; parlor suites, double, $9 and $12 
Make reservations as early as possible to assure satisfactory accommodations. 

Go//, Entertainment 

Passes to local theaters for members, bridge parties and other diversions fo 
the ladies, visits to motion picture studios and to points of interest in and abou 
Hollywood. 

SOCIETY OF MOTION PICTURE ENGINEERS 

HOTEL PENNSYLVANIA, NEW YORK, N. Y. 

(For further details see p. 117) 

Papers for the Spring Convention 

Manuscripts of papers received before April 1st will be given immediate 
consideration by the Papers Committee and the Board of Editors. The best 
of these manuscripts will be selected and given preferred positions upon the 
program of the Convention, with ample time for presentation and discussion, or 
about thirty minutes to one hour. The remaining manuscripts will be considered 
for the program, but with limited time for presentation. 

The remainder of the program will be filled as manuscripts are received, until 
May 1st, after which date no papers will be accepted unless the subject 
matter contained therein is particularly outstanding or timely. Titles and ab- 
stracts of all papers will be published in the May issue if received by April 15th. 



JOURNAL 

OF THE SOCIETY OF 

MOTION PICTURE ENGINEERS 

Volume XXVIII JANUARY, 1937 Number 1 



CONTENTS 

Page 

The Development of the Art and Science of Photography in the 

Twentieth Century C. E. K. MEES 3 

Report of the Standards Committee 21 

Report of the Sound Committee 24 

Report of the Committee on Non -Theatrical Equipment 26 

Report of the Studio Lighting Committee 32 

A Third-Dimensional Effect in Animated Cartoons 

J. E. BURKS 39 

Mercury Arcs of Increased Brightness and Efficiency 

L. J. BUTTOLPH 43 

Standardization of Motion Picture Make-Up M. FACTOR 52 

A Record Word-Spotting Mechanism R. H. HE ACOCK 63 

A Developing Machine for Sensitometric Work 

L. A. JONES, M. E. RUSSELL, and H. R. BEACHAM 73 

Note on the Use of an Automatic Recording Densitometer 

C. N. TUTTLE AND M. E. RUSSELL 99 

Committees of the Society 112 

Society Announcements 1 17 



JOURNAL 

OF THE SOCIETY OF 

MOTION PICTURE ENGINEERS 



SYLVAN HARRIS, EDITOR 

Board of Editors 
J. I. CRABTRBE, Chairman 

O. M. GLUNT A. C. HARDY L. A. JONES 

G. E. MATTHEWS 



Subscription to non-members, $8.00 per annum; to members, $5.00 per annum, 
included in their annual membership dues; single copies, $1.00. A discount 
on subscriptions or single copies of 15 per cent is allowed to accredited agencies. 
Order from the Society of Motion Picture Engineers, Inc., 20th and Northampton 
Sts., Easton, Pa., or Hotel Pennsylvania, New York, N. Y. 
Published monthly at Easton, Pa., by the Society of Motion Picture Engineers. 

Publication Office, 20th & Northampton Sts., Easton, Pa. 
General and Editorial Office, Hotel Pennsylvania, New York, N. Y. 
West-Coast Office, Suite 226, Equitable Bldg., Hollywood, Calif. 
Entered as second class matter January 15, 1930, at the Post Office at Easton, 
Pa., under the Act of March 3, 1879. Copyrighted, 1936, by the Society of 
Motion Picture Engineers, Inc. 

Papers appearing in this Journal may be reprinted, abstracted, or abridged 
provided credit is given to the Journal of the Society of Motion Picture Engineers 
and to the author, or authors, of the papers in question. Exact reference as to 
the volume, number, and page of the Journal must be given. The Society is 
not responsible for statements made by authors. 



OFFICERS OF THE SOCIETY 

President: S. K. WOLF, 250 W. 57th St., New York, N. Y. 
Executive Vice-President: H. G. TASKER, Universal City, Calif. 
Engineering Vice-President: L. A. JONES, Kodak Park, Rochester, N. Y. 
Editorial Vice-President: J. I. CRABTREE, Kodak Park, Rochester, N. Y. 
Financial Vice-President: O. M. GLUNT, 463 West St., New York, N. Y. 
Convention Vice-President: W. C. KUNZMANN, Box 6087, Cleveland, Ohio. 
Secretary: J. FRANK, JR., 90 Gold St., New York, N. Y. 
Treasurer: L. W. DAVEE, 250 W. 57th St., New York, N. Y. 

GOVERNORS 

M. C. BATSEL, Front and Market Sts., Camden, N. J. 
A. S. DICKINSON, 28 W. 44th St., New York, N. Y. 
G. FRIEDL, JR., 250 W. 57th St., New York, N. Y. 
A. N. GOLDSMITH, 444 Madison Ave., New York, N. Y. 
H. GRIFFIN, 90 Gold St., New York, N. Y. 

A. C. HARDY, Massachusetts Institute of Technology, Cambridge, Mass. 
K. F. MORGAN, 7046 Hollywood Blvd., Los Angeles, Calif. 
G. F. RACKETT, 823 N. Seward St., Hollywood, Calif. 
C. H. STONE, 205 W. Wacker Drive, Chicago, 111. 



See p. 112 for Technical Committees 



THE DEVELOPMENT OF THE ART AND SCIENCE OF 
PHOTOGRAPHY IN THE TWENTIETH CENTURY* 

C. E. K. MEES** 

Summary. An account of the developments in practical photography during 
the last thirty-five years and of the progress that has been made in our knowledge 
of the scientific principles of photography. 

Written history has its origin in contemporary documents, docu- 
ments written by those who have taken part in the events described. 
Later, these documents are compared critically by writers who have 
no first-hand knowledge of the events but who summarize the earlier 
writings and prepare descriptions of the events as they appear to 
them at the later date, and the verdict of history is the verdict of 
these writers. It is too early to render a verdict on the progress of 
photography in the first third of the Twentieth Century, but it is 
not too early for one who took part in many of them to describe the 
events. 

An account of this kind is likely to be one-sided. The events 
described in detail are not necessarily those of greatest importance. 
Equally important and interesting events may be ignored or dismissed 
with a brief reference because in them the writer played little or no 
part. However, it is in the hope that this paper may form a contribu- 
tion to the history of photography that it has been written. 

The turn of the century marked a distinct turning point in both 
the science and the art of photography. The earlier scientific workers 
Abney and Hurter and Driifield, in England; Eder and Valenta 
in Austria were ceasing their work, and a new group of students, 
armed with the methods of modern physical chemistry, were just 
beginning to direct their attention to the problems of photographic 
science. In the practice of the art, the field camera with plate equip- 
ment and the print-out processes were being displaced by hand 

* Received October 10, 1936; prepared as a requirement in conjunction with 
the SMPE Progress Award, made to the author at the Semi-Annual Banquet of 
the Society at Rochester, N. Y., October 12, 1936; presented at the Fall, 1936, 
Meeting at Rochester, N. Y. 

** Eastman Kodak Company, Rochester, N. Y. 

3 



4 C. E. K. MEES [J. S. M. P. E. 

cameras using films, from which prints were made on developing-out 
paper by the methods that have persisted ever since. 

The year 1901, therefore, is a suitable date for the beginning of 
a history of the modern era in photography, and it was in that year 
that Dr. S. E. Sheppard and the author started the study of photo- 
graphic science at University College, London. The technical 
methods of photography used at that time depended upon the use 
of gelatin dry plates, mostly of the blue-sensitive type, without any 
dye sensitizing, although a certain proportion of so-called "ortho- 
chromatic" plates sensitized with erythrosine were in general use. 

Most photographers having any pretention to skill and interest 
in their work developed their own plates, although the time when a 
photographer would attempt to make his own plates was already 
long past. Prints were usually made by the printing-out processes, 
which required long exposure to daylight; bromide paper was widely 
used for enlargements. Skilled and enthusiastic photographers made 
many of their prints on platinum paper and on carbon tissue, which 
gave very beautiful prints of great permanence. The carbon process 
and the analogous gum bichromate process enabled a good deal of 
hand control to be introduced into the development of prints and 
were therefore favored by the pictorialists. In the United States 
and Germany, the reigning printing process was the collodio-silver 
chloride paper printed out by daylight and toned with gold, which 
was known in this country as Aristotype. In England the gelatin 
paper was preferred, and trade brands, such as P.O.P. and Solio, 
were used generally. Velox was just coming into use as the first of the 
gaslight papers. It had the advantage that it could be operated with- 
out special precautions in the dimly lighted living rooms of that 
era. 

The cameras generally used were field cameras on tripods, and 
were fitted with rectilinear or convertible anastigmat lenses and 
with mechanical shutters. Almost all the new types of hand cameras 
that were being introduced at the beginning of the century used 
plates. The most elaborate and expensive were the twin-lens cameras 
and the reflexes, of which a great variety were made, especially after 
the higher aperture anastigmats became available soon after 1900. 

There were some very interesting local developments. In France, 
for instance, there was a great vogue for small stereoscopic cameras, 
of very good workmanship, taking plates 45 X 107 mm. in size. The 
best known of these was the Verascope made by Richard, and almost 



Jan., 1937] PHOTOGRAPHY IN TWENTIETH CENTURY 5 

all French amateurs at that time were devotees of stereoscopic 
photography. It is curious that this type of camera did not succeed 
to any great extent in any other country. The reflex camera reigned 
in England, while in Germany portable cameras with focal plane 
shutters, used at eye level, were popular. 

In the first years of the Twentieth Century, Kodaks employing 
film had a wide sale but were regarded by most serious photographers 
as being rather in the toy class. After 1910, however, the modern 
roll-film camera developed rapidly. The second decade, filled as it 
was with the years of war, was characterized by a rapid growth of the 
use of roll films in small and portable cameras, and of professional 
photo-finishing. After 1910, the apparatus and material had been 
largely standardized, as had the finishing procedure, the films being 
developed in deep tanks and printed upon developing-out chloride 
papers, of which Velox was the prototype. The field camera on a 
tripod ceased to be used generally except by the commercial and 
professional photographers. This trend continued to grow until the 
latter half of the third decade, when a new element was introduced 
by the appearance of large numbers of very small, "miniature," 
cameras using film of motion picture width. This revived interest 
in the practice of the photographic processes by the photographer 
himself. 

The shift in general photography through this period, therefore, 
may be summarized as being from the large-size plate camera to the 
portable film camera, with commercialized and mechanized finishing 
methods, and then, to some extent, to the use of miniature cameras, 
with a concomitant emphasis on developing processes and methods 
of enlarging the pictures so as to obtain the best possible definition. 

In this development, an important part was played by improve- 
ments in the methods of sensitizing emulsions to the longer wave- 
lengths of light. Apart from the use of erythrosine for orthochromatic 
plates, practically no color-sensitive materials were manufactured 
before 1900, and, although in 1902 the isocyanine dyes were found to 
be sensitizers, and pinacyanol was made in 1904, the first commercial 
plates employing these new dyes were not produced until 1906. 

The history of the introduction of panchromatic plates may be of 
interest. About 1905, the Neue Photographische Gesellschaft was 
using a process of portrait photography in colors. This involved the 
taking of three separation negatives by means of a mechanically 
operated repeating back, the prints being made upon celluloid films 



6 C. E. K. MEES [J. S. M. P. E. 

coated with dyed gelatin, which was sensitized with bichromate. 
After printing and development by washing out, the three images 
were transferred in register to one support, as in the carbon process. 
This process had been worked out in collaboration with Meister, 
Lucius, and Bruning, later known as the Hoechst Dye Works and now 
a part of the I. G. Farben- Industrie. It was at Hoechst that Dr. 
E. Konig made a number of the isocyanine dyes and that Homolka, 
his colleague, discovered pinacyanol. Konig made the plates for 
the N. P. G. by bathing commercial plates in solutions of pinachrome 
and pinacyanol. The bathed plates did not keep very well, gave a 
good deal of trouble with fog, and were very expensive, but they were 
far more sensitive than any plates that had been made previously. 

In 1906, having taken my examination for the final degree at 
London, I joined the little firm of Wratten & Wainwright in Croydon, 
and the first problem that was referred to me was the preparation 
of plates similar to those that were being made in Germany by 
Meister, Lucius, and Bruning. The English affiliates of the N. P. G., 
the Rotary Photographic Company, like the N. P. G., specialized in 
the manufacture and mechanical printing of bromide paper for the 
preparation of commercial prints and picture postcards. They had 
fitted out a studio in Dover Street, London, for the practice of com- 
mercial color portraiture. For this work, they needed plates; and 
as the German plates gave some trouble and were very expensive, 
they desired to obtain plates made in England, and the English agent 
of Meister, Lucius, and Bruning approached Wratten & Wainwright. 
The idea was that the plates should be made by bathing, following 
the methods worked out by Konig. I had had a good deal of ex- 
perience with the difficulties attending the preparation of bathed 
plates, so that it seemed to me that it would be better to attempt to 
coat a sensitized emulsion. 

It is now rather hard to realize how difficult it was to start the 
coating of panchromatic plates without any experience in such a 
thing. We had no safelights, of course, and we had to get the coating 
machine running by red light, so that the emulsion covered the plates 
properly. Then we turned out the light as the plates were running 
through, and pushed over to the side of the maclu'ne the plates coated 
in red light, so that the girl at the other end of the coating machine 
would know that they had been fogged and would put them on a 
separate rack. When the first plates coated in total darkness came to 
her, she started to pick them off the machine and put them into racks 



Jan., 1937] PHOTOGRAPHY IN TWENTIETH CENTURY 7 

to dry. All went well for a few minutes. Then crash went a plate 
on the floor, and the girl gave a squeal. Then crash went another, 
and she gave another squeal ; and when a third plate did the same, her 
nerves gave way, and she curled up on the floor in hysterics while 
the machine showered her with sticky plates. Wratten* shot through 
the pitch darkness straight toward those hysterical squeals, shoved 
their producer ruthlessly aside, and collected the precious plates. 
The next morning the plates were dry. My laboratory was at home, 
seven miles off, so I came down on my motorbike, picked up a few 
plates, took them back and tested them. They were fine, so I shot 
back to tell Wratten the good news, and the next day took the plates 
up to London and sold them. 

One of the troubles was inspecting the plates. We had to make 
sure that they were covered with emulsion, so we put a candle at the 
end of a long passage and found out by trial that when our eyes were 
rested we could see the plates by the candle light and that it took 
five or six seconds to produce a visible fog on them. We inspected 
the plates by picking up one from a pile in darkness on one side of 
the passage and transferring it rapidly to darkness on the other side, 
looking at it in transit. That insured that the plates would not be 
left in the light. Within a few weeks, of course, we had designed 
green safelights and proper coating and inspection methods, and the 
technic of panchromatic material making was well on its way. 

I am afraid that the first plates supplied to the Rotary Photo- 
graphic Company were labelled bathed plates. Photographers were 
very conservative then, as now, and although the emulsion sensitized 
plates were better than the German plates, the users would not have 
touched them if they had known how they were made. Within 
three months, however, we found methods for combining the pina- 
chrome and pinacyanol in one emulsion, and made the Wratten 
panchromatic plate which has been on the market ever since. 

Soon after we made the panchromatic plates, the question of light 
niters came up. Not many commercial light filters were made at that 
time. A few were made commercially in England and Germany, 
and enthusiasts prepared filters for themselves. Freiherr von Hiibl, 
Dr. Konig, and Professor Miethe had published formulas for pre- 
paring them. Most of the yellow filters used were made of brownish 
yellow glass having very poor correcting power, and when I got to 
Wratten & Wainwright I found that we were selling filters of that 

* S. H. Wratten, son of the founder of the firm, F. C. L. Wratten. 



8 C. E. K. MEES [J. S. M. P. E. 

type. Mr. Wratten, however, was experimenting with filters made 
of gelatin by coating upon a polished glass plate a thick solution of 
gelatin dyed with tartrazine and then, after drying the plate, stripping 
off the gelatin. This was an excellent scheme for supplying cheap 
filters, and, in a short time, we were marketing both the gelatin 
filters and the same filters cemented between glass plates. 

Early in 1907, Novak published an account in Photographische 
Korrespondenz of a new dye, filter yellow K, developed at the Hoechst 
Dye Works, which was greatly superior to tartrazine for use in 
filters. I read the article when it appeared and cabled to Dr. Konig 
for a sample of the dye, which arrived on a Friday evening. A few 
tests showed that it was all that Novak had claimed for it, and 
Wratten and I worked through the week-end, first testing the dye 
and measuring its properties, and then preparing coatings of gelatin 
filters and examining them. When the British Journal of Photography 
appeared on Thursday that week, with an article on the new dye, 
it also carried the Wratten advertisement of a series of Wratten K 
filters K-l, K-2, and K-3 which have since been famous in photo- 
graphic circles. A set of gelatin tricolor filters had been made in 
1906, and after the K filters were made, a very considerable proportion 
of the present Wratten filters was worked out and placed upon the 
market, including filters for all sorts of scientific work, such as 
photomicrography and spectroscopy. 

At that time darkrooms were generally lighted by simple oil lamps 
fitted with red or yellow glass of very doubtful safety and efficiency. 
As a result of a study of the properties of darkroom safelight filters 
in relation both to vision and to the effect produced upon materials 
exposed to them, a series of darkroom lamps and of safelight filters 
were introduced that proved entirely satisfactory and are now uni- 
versally adopted. 

In 1908, the first process panchromatic plates were made, and the 
English photoengravers adopted these for making three-color half- 
tone separations by the direct process instead of using collodion 
emulsion. It is curious that collodion emulsion maintained its position 
against the panchromatic dry plates for a much longer period in the 
United States than it did in England. 

Panchromatic plates were used first by the commercial photogra- 
phers for the photography of colored objects paintings, furniture, 
and so forth and also by enthusiastic amateurs who liked to get 
correct color rendering in their pictures. During the War, pan- 






Jan., 1937] PHOTOGRAPHY IN TWENTIETH CENTURY 9 

chromatic plates were used with color filters for photography from 
the air, because through haze they gave very much better pictures 
than those taken on the non-color-sensitive materials. After the 
War panchromatic plates were firmly established, and they began 
to find wider use by both amateur and professional photographers. 

The first commercial panchromatic film was made as early as 1914 
for use in color motion picture processes, but was not used generally 
for motion picture work until 1925. Panchromatic motion picture 
negative film was then introduced generally throughout the trade and 
displaced the older orthochromatic film very rapidly, the change 
being accelerated in 1928 by the adoption of sound recording and of 
tungsten lamps in the place of arc lamps in the studios. This film, 
however, like all those before it, was still sensitized with the pina- 
cyanol dye introduced as long ago as 1904. Although a certain 
amount of research on sensitizing dyes had been carried on in the 
manufacturers' laboratories, nothing really new had been produced 
until in 1928 it was realized that the thiocarbocyanines that had been 
studied by W. H. Mills and F. M. Hamer represented a type of dye 
having great advantages over those previously known. L. G. S. 
Brooker was successful in making carbocyanines from naphtha- 
thiazole instead of from benzothiazole, which had been used by Mills 
and Hamer, and these dyes sensitized very strongly for the red, so 
that for the first time sensitizers for the red had been produced that 
were superior to pinacyanol. 

Whereas the earlier dyes tended to decrease the general speed of 
the emulsion and to give a certain amount of fog, the new dyes in- 
creased the total speed and at the same time made the emulsion re- 
sistant to fog. It was possible, therefore, to prepare a new type of 
panchromatic film, which became known as "supersensitive" film. 
This at once affected the methods used for the production of motion 
pictures in the studio: it gave the photographer a much simpler 
control of his lighting and made possible the development of studio 
photography to a previously unheard of degree. 

The same type of emulsion is used for making films for portraiture, 
and films and plates for commercial photography and three-color 
work, so that with the very general use of panchromatic materials 
in the last few years, the art of photography has entered a new phase. 

The introduction of the new dyes produced a change in photo- 
graphic practice in another direction: they increase the general 
sensitivity of slow, fine-grain emulsions to a much greater extent than 



10 C. E. K. MEES [J. S. M. P. E . 

they do that of high-speed emulsions, so that while a fine-grain 
emulsion, even after sensitizing, is slower than a sensitized emulsion 
having grain of average size, the two are much more nearly alike in 
speed than before sensitizing. This made it possible to prepare 
films of satisfactory effective speed of which the image structure 
allows a considerable degree of enlargement, and this, in turn, pro- 
duced satisfactory results from miniature cameras taking very small 
pictures. As a result, there has been a great rise in the popularity 
of cameras taking pictures about an inch by an inch and a half in 
size. The portability and mechanical excellence of these cameras 
make them attractive, while the fine-grain films available for their 
use give results comparable with those attainable with the larger 
instruments. 

The rise of amateur cinematography dates from 1920, at which 
time a number of experimenters were making small cameras, generally 
using 35-mm. film slit down the center, which gave pictures a quarter 
of the standard size. No general use was possible, however, unless 
facilities were available for having the film processed speedily and 
effectively. The motion picture laboratories could not be expected 
to devote to the short lengths of amateur film the attention bestowed 
upon the commercial film, of which millions of feet are processed 
every week; and the general public did not know where to send the 
film. The use of motion pictures by amateurs, therefore, was at 
first confined to a few enthusiasts. 

One of the difficulties with the ordinary negative and positive 
process is that making a single print of high quality is quite difficult. 
In professional work, the rush prints often require reprinting, and it 
is only after a negative has been assembled and tested carefully 
throughout that prints of high quality can be made at a low cost. 
Moreover, the use of smaller pictures on standard film involved more 
evident graininess than that shown by the standard pictures. 

All these problems were solved by the introduction of the reversal 
process. The reversal of a photographic image by the use of potassium 
permanganate as a solvent for the developed silver, and the redevelop- 
ment of the positive, was introduced by Namias. The process was 
greatly improved by the suggestion of Capstaff that the re-exposure 
after bleaching should be controlled in amount, which made it possible 
to use thickly coated material and, at the same time, to correct for 
differences in the initial exposure in the way that such differences 
are compensated for in ordinary printing from the negative. 



Jan., 1937] PHOTOGRAPHY IN TWENTIETH CENTURY 11 

The reversal process with controlled second exposure was an in- 
stant success and, as a result, processing stations in which the amateur 
could have his film processed were installed all over the world. 
There is little doubt that it was this development that determined 
the success of the 16-mm. program of amateur cinematography. 

A later improvement was the use of automatic compensation in the 
second exposure by the use of photocells, which operate so quickly 
that change occurs within a few frames on 16-mm. film. The de- 
velopment of portable, spring-driven cameras and of extremely con- 
venient and reliable projectors has been very remarkable, and ama- 
teur cinematography is now an important division of the photographic 
industry. 

This brief account of the development of general photography 
during the first third of the Twentieth Century is, of course, very 
incomplete. Nothing has been said of the advances made in photo- 
graphic optics, in which progress has been continuous, and only a 
passing mention has been made of the introduction of sound record- 
ing, which has revolutionized the production of motion pictures. 

Many of the great applications of photography have, perforce, 
been ignored. Radiology, aerial photography, documentary photog- 
raphy, and the whole field of the graphic arts would each require 
a separate article to do justice to them. But one division of the 
photographic art is of so distinct a nature, and holds so much promise 
for the future, that a section of this paper must be devoted to it. 
This is the field of photography in color. 

While rapid and continuous advance has characterized the art 
of photography during the first third of the Twentieth Century, this 
can not be said of color photography, although at the present time 
there is great activity in that field. The many processes of color 
photography that have hitherto been introduced, both experimentally 
and commercially, must be regarded as the early and, on the whole, 
unsuccessful forerunners of the processes that will eventually succeed. 
The nature of the really practical processes of color photography 
are probably only just beginning to be understood. 

Before 1900, indeed long before, the various additive and subtrac- 
tive processes of color photography had been outlined clearly, and 
attempt after attempt was made to realize them in practice. In the 
first years of the Twentieth Century, there was much activity in 
the design of one-exposure, three-color cameras intended for portrai- 
ture and commercial work. The negatives were made in special 



12 C. E. K. MEES [J. S. M. p. E. 

cameras or in ordinary cameras fitted with repeating backs. The 
prints were usually made by some modification of the carbon process, 
depending upon the sensitizing of gelatin with bichromate and the 
superposition of the three dye layers so obtained. The process used 
by the Neue Photographische Gesellschaft in 1905 has already been 
mentioned. At about the same time, the Hoechst Dye Works worked 
out the pinatype process and another process, using the leuco bases 
of dyes, which was termed "pinachromy." More recently, similar 
processes have been placed upon the market as "Carbro" and "Wash- 
off Relief" processes, all of them depending upon the printing of the 
three images in dyes and the superposition of the dyes in one image. 

The screen-plate processes, which had been attempted commercially 
by Joly and McDonough, reached the apex of success in the Lumi&re 
Autochrome process, in which the filter units were composed of dyed 
grains of rice starch, and with which beautiful color transparencies 
can still be obtained. Other screen-plate processes depended upon 
the use of dyed particles of resin or of regular screens made by 
printing bichromated fish glue or albumen. An application of the 
screen-unit process to film is provided by the Dufaycolor process, 
in which resist lines are printed mechanically upon a dyed cellulose 
base, so that the units can be bleached and dyed in different colors. 

In the field of color cinematography, the first practical process 
was Kinemacolor, a two-color additive process, in which the unit 
pictures were taken successively through red and green filters and 
projected in the same manner. Numberless minor variants on this 
theme have been protected by patents and exploited on an experimen- 
tal commercial basis. An attempt at a complete three-color additive 
process was launched by Gaumont as early as 1914, but the difficulties 
in projection proved fatal. In all processes that require a special 
projector, it has been a formidable difficulty that it is hard to per- 
suade exhibitors to provide the projection equipment until they are 
assured of a continuous supply of films, and that it is impossible 
for producers to provide films until a sufficient number of theaters 
are equipped to project them. 

About 1915, there was much activity in the field of subtractive 
two-color motion picture processes. Two negatives are taken either 
through two lenses by means of a beam-splitter or a bipack system, 
and these are printed upon opposite sides of double-coated film, one 
side being transformed by chemical treatment into a green image and 
the other into a red image. So active were the patentees in this 



Jan., 1937] PHOTOGRAPHY IN TWENTIETH CENTURY 13 

field that two patents covering films of this type issued independently 
from the United States office within a few weeks of each other, but 
satisfactory arrangements having been made with the patentees, 
the development of this work went forward on a small-scale, com- 
mercial basis. 

The ingenuity shown in overcoming the practical difficulties of 
making color pictures has been quite remarkable. The Lumire 
Autochrome process was a triumph of technology. It seems quite 
impossible that dyed starch grains of so small a size could be made into 
a layer of colored filters with the precision shown in the Lumie"re 
plates. The earliest Technicolor process, again, in which thin films 
carrying wash-off relief images were cemented in register back to 
back was a triumph of technical skill; while the imbibition process 
of the Technicolor Company represented the first really practicable 
process for the production of motion pictures in large quantities at 
a cost at which they could be sold. In this process, the wash-off 
reliefs or matrices are used to form dye images, which are transferred 
by imbibition to a gelatin layer, and after this operation was carried 
out successfully with two colors, the Technicolor Company was able 
to produce three-color images, which today represent the furthest 
commercial development in the production of color motion pictures. 

One other process, which is distinguished both by the simplicity 
of its operation and the beauty of its results, is that invented by 
Berthon in 1908, in which microscopic embossed lenses are used to 
form the optical equivalent of screen units both in taking and in 
projection, so that a color picture is produced by the simple use of a 
triple filter on the lens and the special lenticularly embossed film. 
This process is particularly convenient for amateur cinematography 
since it is extremely suitable for use with the reversal process. Its 
development in the professional field is quite feasible, but still presents 
the difficulty that some modification is required in the projection 
machines and that special precautions must be taken in view of the 
loss of light occasioned by the filters. 

Throughout this development of color photography, a silent but 
very important role has been played by the emulsion maker, who has 
provided the special sensitizing required for all the color processes in 
turn. The new prospects that are opening up to us in the field of 
color photography are rendered possible to a very great extent by 
the discovery of the new sensitizers referred to previously. 

From time to time, it has been suggested that instead of using 



14 C. E. K. MEES [J. S. M. P. E. 

three separate negatives and then superimposing the prints, color 
photographs might be taken on specially prepared multiple layers of 
emulsion each one sensitive to one of the primary colors. The 
colors are separated in the depth of the film. After exposure, each 
layer produces a colored image, so that a subtractive picture is ob- 
tained directly. 

An even more attractive suggestion is the mixed-grain process, 
in which it is suggested that the silver bromide grains themselves be 
differentiated in sensitivity and in their capacity for forming a final 
colored image upon being developed, so that grains that were sensitive 
to red would become blue, and grains in the same layer sensitive to 
green or blue would become respectively magenta and yellow. Up 
to the present, no one has been able to realize this last suggestion, 
but workers in several directions have been able to realize the multi- 
pack processes of color photography. 

In the early part of 1935, a multicoated film was placed upon the 
market under the name of "Kodachrome" and has been successful 
for amateur cinematography. It is possible to take 16-mm. or 8-mm. 
pictures in color with no more complication or difficulty than if they 
were taken in black and white. Bela Caspar has succeeded in realizing 
another multilayer process, in which the light-sensitive layers of 
emulsion contain the dyes that form the final images, and the dye 
is destroyed by chemical treatment in proportion to the silver image 
produced in development as a result of exposure. 

The multilayer processes are developing very rapidly, and it 
would appear that we are on the verge of an era of great activity in 
color photography. 

When we turn from the art of photography to its science, the prog- 
ress made in the Twentieth Century as compared with what had 
gone before is very significant. 

The chief phenomena of photography were discovered in the Nine- 
teenth Century. Abney had observed the failure of the reciprocity 
law, and the intermittency effect; he had also studied the color- 
sensitivity of the silver salts; Vogel discovered the phenomenon of 
optical sensitizing; and Eder had investigated the sensitizing power 
of a very large number of dyes. The earliest work on sensitometry was 
due to Abney, who studied the relation between the transparency of 
a photographic image and the exposure; and a great deal of work 
had been done on the reactions of the latent image, so that the 
literature contains much discussion of the structure of the latent 



Jan., 1937] PHOTOGRAPHY IN TWENTIETH CENTURY 15 

image and of its chemical reactions. The simple chemistry of de- 
velopment had been studied, and the types of organic developing 
agents defined, but very little was known of the chemistry of the 
development process, a matter that is still to some extent obscure. 

By far the most important work on the theory of photography was 
that due to Ferdinand Hurter and V. C. Driffield, who in 1890 
published their "Photo -chemical Investigations," in which they laid 
the foundations of sensitometry and, at the same time, outlined the 
general theory of the photographic process. This was followed in 
1898 by a second paper, dealing largely with the chemistry of develop- 
ment, and by a series of shorter articles in which they discussed the 
general principles of tone reproduction. 

Hurter and Driffield were amateur photographers who became 
interested in the application of scientific methods to practical pho- 
tography. Hurter was a competent physical chemist engaged in 
the manufacture of alkali, and Driffield was an engineer in the same 
business and had a thorough grasp of mathematical methods. They 
desired to place the exposure of their pictures on a scientific basis, 
and to that end they designed a chemical photometer by means of 
which they could measure the variation of sunlight throughout the 
day and throughout the year. Having obtained this information, 
they constructed an exposure calculator, which they termed the 
"Actinograph," and at this point they found that they needed in- 
formation with regard to the intrinsic sensitivity of the photographic 
plates that they were using. They then started their photochemical 
investigations. They exposed plates to known quantities of light, 
expressed in candle-meter-seconds, developed them under controlled 
conditions, and measured the opacity of the developed image on a 
photometer of their own devising. They showed that the logarithm 
of the opacity, which they termed the "density," is proportional to 
the mass of silver obtained. They then plotted the density obtained 
after development against the logarithm of the exposure given, get- 
ting a curve which they termed the "characteristic" curve of the 
material. They showed that as development is increased, the straight- 
line portion of the curve rotates about a point upon the exposure axis, 
which they termed the "inertia point." They used this inertia as a 
means for calculating the now well known "H&D sensitivity" of 
the materials, which they expressed as the value 34/i in order to get 
a convenient factor for use with their actinograph. They termed the 
slope of the characteristic curve 7, and showed that density increases 



I6 C. E. K. MEES [J. S. M. P. E. 

during development as an exponential function of the time, finally 
reaching a limiting density. 

When Dr. Sheppard and I took up this subject in 1901, advances 
could be made in two directions: better apparatus than that used 
by Hurter and Driflield could be designed, and the development of 
physical chemistry made it possible to study the statics and dynamics 
of the development much more thoroughly than had been possible 
for Hurter and Drifiield. Hurter and Driffield's light-source was at 
first an actual spermaceti candle, and later a pentane standard lamp. 
This was replaced by acetylene burners in which a screened portion of 
the flame gives a convenient, uniform light-source, which is much 
more convenient. A thermostat was designed to control the tempera- 
ture of development, so that a long series of experiments could be 
made without any danger of a variation in temperature. The primi- 
tive H&D sector-wheel was placed by a better made instrument in 
which the light-source was enclosed and the plates could be inserted 
in plate holders, while the whole instrument was used in an ordinary 
room. Various photometers were employed in place of the primitive 
instrument designed by Hurter and Driffield. The Hiifner spectro- 
photometer was adopted for a considerable time and gave satisfactory 
results although it was unnecessarily complicated and very expensive. 

These improvements in apparatus were of considerable importance 
in the development of sensitometry. The instruments designed at 
that time are now obsolete, but at no period since has any change in 
the apparatus used in sensitometry been comparable in importance 
with that made at the beginning of the century. Our repetition of 
Hurter and Driffield's work on the relation between the density and 
the mass of silver, the shape of the characteristic curve, the effect 
of development upon density, etc., merely extended and verified the 
conclusions to which the two pioneers had come in the course of 
their work. The increase in 7 during development was correlated 
with Hurter and Driffield's work on density, and the importance 
of y * was established. Our first paper was published as early as 
1903 and dealt with the development factor (7). 

This was followed by a series of publications dealing with the 
physical chemistry of development and with the microscopic struc- 
ture of the image. The work on development was a fundamental 
study of the chemical statics and dynamics of the reactions both in 
the ferrous oxalate developer, which had been used by Hurter and 
Driffield, and in the alkaline developers, especially hydroquinone. 



Jan., 1937] PHOTOGRAPHY IN TWENTIETH CENTURY 17 

Another series of investigations established the nature of the in- 
organic desensitizers* and of the reactions that destroy the latent 
image, the whole of this work being collected and published in book 
form. 

This first collaboration of Dr. Sheppard and myself ended in 1906, 
when I joined Wratten & Wainwright and started work on pan- 
chromatic plates and light filters. The introduction of panchromatic 
plates and the marketing in 1907 by the Lumiere Company of the 
first direct color-plate, the Lumiere Autochrome, directed general 
attention to the subject of color photography, and the chemistry 
of the photographic process was rather neglected for some years. 
There were, however, a number of contributions relating to the 
physics of photography. In 1909, I published the first paper on pho- 
tographic resolving power, and this was followed by articles by 
several workers, notably by Dr. Scheffer. A. Callier made systematic 
measurements of the scattering power of photographic deposits and 
established the conditions that should be used for the measurement 
of photographic densities. 

When the research laboratory of the Eastman Kodak Company 
was established in 1912, the immediate problems of photographic 
theory could be divided into four main sections : 

The first was the theory of tone reproduction the study of how 
closely a photographic reproduction resembles the original in the 
distribution of brightness, and the prediction from the sensitometric 
characteristics of photographic materials of their ability to reproduce 
tone values with more or less correctness. The first step required in 
this work was an investigation of the sensitometry of printing papers. 
The results were published in the form of a paper in 1914 and pre- 
pared the way for a general study of tone reproduction. This was 
undertaken at about the same time by workers in Germany, England, 
and the United States. Dr. Goldberg in Germany, F. F. Renwick in 
England, and L. A. Jones in the United States published a series of 
papers that completed the work that Hurter and Driffield had 
started in 1890 clearing up the nature of the photographic process 
from the standpoint of physical optics. It is now standard practice 



* Desensitizers of another kind were discovered later by Luppo-Cramer, who 
found that dyes such as phenosafranine are powerful desensitizers and can be used 
for the treatment of an emulsion before development, so that more light is possible 
in the darkroom. 



18 C. E. K. MEES [J. S. M. p. E. 

to compute the reproduction that will be obtained from any subject 
by using the characteristic curves of the materials. This solution of 
the problem of tone reproduction has been particularly valuable in 
connection with sound recording, and promises to be of even greater 
importance in connection with the problems of color photography. 

Another group of photographic investigations is that which deals 
with the structure of the photographic image, involving the whole 
question of resolving power, the sharpness of images, and the distor- 
tion of the position and shape of small images. A considerable 
series of papers, largely from the Kodak research laboratories, seems 
to have placed these problems on a firm footing, so that the general 
physics of photography may be said to be well understood. 

A part of the physics of photography that is of great importance 
in the applications of photography to scientific work, such as spec- 
troscopy and astronomy, is the relation between the exposure pro- 
duced and the intensity of the light. It was shown by Abney more 
than fifty years ago that the image produced is not independent of 
the intensity, and that if a photographic material is exposed to a 
weak source for a long time, the image produced is weaker for the 
same amount of energy than if the light were more intense and the 
exposure shorter. This has always been known as "the failure of the 
reciprocity law" since Bunsen and Roscoe stated that for photochemical 
reactions, time and intensity were interchangeable and that the same 
amount of energy produced the same result. The failure of the 
reciprocity law has been a favorite subject for photographic investiga- 
tors since Abney's discovery. It was studied exhaustively by Eng- 
lisch at the beginning of the Twentieth Century, but measurements 
of the necessary precision were first made in the Kodak laboratories 
soon after their establishment. Our knowledge of the subject has 
recently been improved very much by the work of J. H. Webb, who 
showed that Abney's intermittency effect can be explained in the 
terms of the reciprocity failure, and that the reciprocity failure 
depends upon the temperature at the time of exposure. The reverse 
action, which produces the failure, has apparently a high temperature 
coefficient and is therefore presumably of a chemical nature. It is 
probable that the full explanation of the reciprocity failure will come 
when we understand the nature of photographic exposure itself. 

Far less progress has been made in dealing with the chemistry of 
photographic development. After the work by Sheppard, the 
chemical reactions in an alkaline developer were studied by a number 



Jan., 1937] PHOTOGRAPHY IN TWENTIETH CENTURY 19 

of German workers, notably by Luther and his students. Frary and 
Nietz attempted to measure directly the reduction potential of 
developers, and this work was continued by Nietz in the Kodak 
laboratories, who carried on an extended series of investigations on 
the physical chemistry of development. Quite recently, Reinders 
and Beukars in Holland have published some important work on 
this subject, and this has been followed up in the Kodak laboratories, 
so that it is quite possible that in the near future we may attain to a 
satisfactory understanding of the general chemistry of development. 

The classical problem of photographic science is, of course, the 
phenomenon of exposure and the nature of the latent image produced, 
and to this must be added the structure and properties of the light- 
sensitive silver compound. This problem was attacked intensively 
in the years immediately folio wing the War, both in the Kodak labora- 
tories and by the British Photographic Research Association in 
England. The first work, however, was published by Dr. T. Sved- 
berg, of Sweden, who made a statistical study of the number of grains 
that became developable after exposure. Sheppard and Trivelli in 
the Kodak laboratories engaged in a laborious and difficult analysis of 
the distribution of grain sizes in emulsions, and showed that the 
sensitivity was associated with the size of the grains and that the 
distribution of different sizes of grains conditions the photographic 
properties of the emulsion, a possibility that I had suggested some 
years before. It was already known that development proceeded 
from specific centers in the grains, and the English work at this time 
showed that the centers were in all probability specific specks of some 
substance, other than silver halide, present on the grain before ex- 
posure and acting as sensitizing specks. 

For many years, it had been known that photographic gelatin 
differed very much in its ability to produce sensitive emulsions, and 
it was suspected that this was due to some small impurity. As the 
result of a series of brilliant investigations, Sheppard showed in 
1925 that the substance in gelatin that is responsible for sensitizing 
contains sulfur in a labile form, and that in all probability the sensitiz- 
ing specks consist of silver sulfide. 

Following this, there was a very active discussion, in which all 
countries took part, as to the phenomenon of exposure, and a number 
of theories were advanced as to the action of light upon the silver 
bromide grains. 

In connection with this subject, it should be mentioned that the 



20 C. E. K. MEES 

investigations of Sheppard and Vanselow showed that the first 
effect of light was to liberate an electron from the bromide ion of 
the silver bromide crystal; this electron then united with a silver 
ion to form an atom of silver. At about the same time Toy and 
Harrison showed that the wavelength of the light that produced 
photoconductivity in silver bromide could be reconciled with the 
spectral sensitivity of the material. The recent work of Hilsch and 
Pohl on the primary photoelectric effect is in agreement with the 
earlier results, and seems to be on the point of yielding a simple and 
clear theory of the action of light upon the photographic material 
and of the nature of the exposed grain. While the investigations on 
grain distribution and Sheppard's work on sensitizing have added 
greatly to our knowledge of the structure of silver halide emulsions, 
and while the properties of the silver halide grains themselves were 
clarified by the work of Sheppard and Trivelli, who published a 
monograph on the subject, the actual properties of photographic 
emulsions and their relation to the methods of preparation are still 
quite obscure, and it will undoubtedly require many years before 
these matters are cleared up. 

In the last thirty-five years, there has been developed a coherent 
theory of the photographic process. At the beginning of the century, 
the gaps in our knowledge were so many and so deep that a science 
of photography could scarcely be said to exist. Today, many of 
those gaps have been filled and the structure of photographic science 
is continuous. The further development of that structure is a task 
for the future a task full of promise and of interest. 

DISCUSSION 

MR. RICHARDSON: How was the graininess of film reduced? 

DR. MEES: The main factor in diminishing the graininess of negative mate- 
rials was the use of sensitizers to increase the speed of the finer-grained emulsions. 
As I mentioned, the sensitizers increase the speed of fine-grained emulsions to a 
greater extent than that of coarse-grained emulsions. I do not mean that they 
get up to the same speed finally. As a general rule, a coarse-grained emulsion 
when sensitized is faster than a fine-grained one, but the difference is much less 
when sensitized than before sensitization. By making finer-grained emulsions 
and then sensitizing them, it was possible to obtain sufficient speed without unde- 
sirable graininess for the motion picture screen. 






REPORT OF THE STANDARDS COMMITTEE* 

During the spring and early summer a series of meetings have been 
held dealing largely with the details of the new drawings prepared 
for the revised standards under the direction of Mr. G. Friedl, Jr. 

These drawings have been designed to fit better into the plan of 
the American Standards Association, and it is hoped that they will 
serve as models for future drawings so as to avoid misunderstandings 
similar to those encountered in connection with the 16-mm. sound- 
film standards. 

The serial numbers used in the past for the Standards drawings 
have been given up, and it is proposed to use a classification such 
that the drawing of a given standard will retain its number through 
any series of revisions by changing only the final number, indicating 
the number of the revision. It is believed that the new numbering 
will be useful not only to those consulting the charts, but to the 
Standards Committee itself, inasmuch as the numbering system 
will show any standards that are missing or have been neglected. 
Each drawing will contain the American Standards Association num- 
ber as well as the SMPE number. 

The question of the single type of perforation adopted by the 
Society for both positive and negative is being studied intensively 
by a subcommittee under Mr. J. A. Dubray, Chairman of the West 
Coast Branch of the Standards Committee. Although no final 
report has been prepared, it appears that the difficulties involved 
in changing to the standard SMPE perforation for the negative are 
very considerable, chiefly because of the background negatives being 
used at the present time. Mr. Dubray is again going into the ques- 
tion of changing the longitudinal dimension of the positive type of 
perforation, reducing it from 0.078 to 0.073 inch, so that the film 
will fit upon positioning pins and other apparatus designed for the 
old Bell & Howell perforation. 

The proposal has been received from the German Standards As- 
sociation that 16-mm. sound-film spools be standardized with square 
holes on each side instead of with one square and one round hole, 

* Presented at the Fall, 1936, Meeting at Rochester, N. Y. 

21 



22 REPORT OF STANDARDS COMMITTEE [J. S. M. P. E. 

as is common practice in this country to prevent the amateur pro- 
jectionist from putting the spool on backward. This proposal has 
been referred to the Non-Theatrical Equipment Committee, which 
has made careful inquiry among the various manufacturers and 
has received opinions from most of them. There has been no meet- 
ing of the Standards Committee since the report of the Non-Theatrical 
Equipment Commitee was received, and, therefore, no action has 
been taken upon it. Standardization of the 2000-ft. reels has been 
withheld pending further recommendations from the Exchange Prac- 
tice and Laboratory Practice Committees. 

Two items have been discussed at considerable length at many of 
the Standards Committee meetings, one in regard to the standardiza- 
tion of sprockets. Many members feel that the Committee should 
not standardize sprockets of any sort but that their design should be 
left to the projector and camera designers to achieve the best results 
with standard film. The second question has to do with better 
methods of obtaining complete uniformity in standardization through- 
out the world. Even when the intention of the various standards 
committees is the same, small differences invariably occur, sometimes 
due to the fact that tolerances chosen in one country are different 
from those in another; and sometimes in this country the tendency 
is to round off figures in inches, whereas in France and Germany 
the figures are rounded off in terms of millimeters. No satisfactory 
conclusions have been reached in either of these matters. 

Requests have been sent to various manufacturers for data on 
8-mm. film, sprockets, and film gates in order that standards may be 
drawn up for 8-mm. equipment. 

The question of a standard reduction ratio for reducing 35-mm. 
film to 16-mm. film has been discussed, and the Standards Committee 
has requested Mr. Friedl to prepare a paper on the subject to be 
presented at the Fall, 1936, Meeting at Rochester, in order to evoke 
discussion from the members of the Society. The question is fairly 
complicated, and necessarily involves compromises of some sort, 
since the shapes of the pictures are not identical. Members of the 
Society having special interest in or knowledge of the question are 
requested to send their opinions to the Standards Committee. 

E. K. CARVER, Chairman 

P. ARNOLD L. N. BUSCH A. COTTET 

F. C. BADGLBY W. H. CARSON L. DE FEO 

M. C. BATSEL A. CHORINE A. C. DOWNES 



Jan., 1937] REPORT OF STANDARDS COMMITTEE 23 

J. A. DUBRAY E. HUSE C. N. REIPSTECK 

P. H. EVANS C. L. LOOTENS H. RUBIN 

R. E. FARNHAM W. A. MACNAIR O. SANDVIK 

C. L. FARRAND K. F. MORGAN H. B. SANTEE 

G. FRIEDL, JR. T. NAGASB J. L. SPENCE 

H. GRIFFIN N. F. OAKLEY A. G. WISE 

R. C. HUBBARD G. F. RACKETT I. D. WRATTEN 
W. B. RAYTON 

DISCUSSION 

MR. PALMER: One of the new drawings specifies 26 frames between the sound 
aperture and the picture aperture. I thought it was 25 frames. 

DR. CARVER: Twenty-five is the present standard. The standard in Europe 
is 27, and the Standards Committee is proposing that we compromise on 26. 

It seems only fan* that, since the European Committees have generously 
agreed to adopt our standard in regard to the side of the film for the 
sound- track, we should meet them half-way in respect to this item. I think 
that the change from 25 to 26 is advisable. The Standards Committee is unani- 
mous on this point. 



REPORT OF THE SOUND COMMITTEE* 



In one of the projects upon which the Sound Committee has been 
working considerable progress has been made namely, the elimina- 
tion, through the use of a Standard Frequency Reference film, of 
the wide differences in recording characteristics that have occurred 
in the past. As stated in the last report of the Sound Committee, 1 
a Primary Frequency Reference Standard has been made, and 
twelve Secondary Standards made and calibrated in terms of the 
Primary Standard. The calibrations of the Secondary Standards 
were given in the previous report. 

The Secondary Frequency Reference Standards have been sent to 
the various studios on the East and West Coasts, and, through the 
courtesy of the heads of the various studio sound departments and 
with the cooperation of the Academy of Motion Picture Arts and 
Sciences, the recording frequency characteristics of the various studios 
have been determined and referred to this Frequency Reference film, 
and the results from the various studios have all been plotted upon 
the same sheet of paper. This is the first time, to our knowledge, 
that comparable data from the various studios have been assembled 
into one set of curves. The results bring to the attention of recording 
directors in a forceful, graphic, and quantitative form the reason for 
the large variation in the quality of the sound-track from the various 
studios when reproduced in the theater. 

It was hoped that it would be possible to present copies of these 
curves at this time, but a study of the results seems to arouse con- 
siderable doubt as to whether certain studios followed precisely the 
procedure laid down by the Sound Committee in obtaining the data 
at present available. Consequently, it is the opinion of both the 
SMPE Sound Committee and that of the Academy of Motion Picture 
Arts and Sciences that publication of the data should be withheld 
until such time as its accuracy has been checked carefully. It is 
safe to say, however, that the data collected have already given 

* Received October 12, 1936; presented at the Fall, 1936, Meeting at 
Rochester, N. Y. 

24 



REPORT OF SOUND COMMITTEE 25 

great impetus to the desire for greater uniformity in the sound record- 
ing characteristics of the various studios. 

The Primary Frequency Reference Standard, originally measured 
in 1935, is being remeasured this year, in order to determine the ex- 
tent to which aging of the film has affected the level of the sounds 
recorded. Microdensi tome trie measurements made by the Eastman 
Kodak Company do not show deviations in excess of the experimental 
error that may reasonably be expected. 



P. H. EVANS, Chairman 

M. C. BATSEL K. F. MORGAN R. O. STROCK 

L. E. CLARK O. SANDVIK H. G. TASKER 

F. J. GRIGNON E. I. SPONABLE S. K. WOLF 

REFERENCE 

1 Report of the Sound Committee, J. Soc. Mot. Pict. Eng., XXVI (Jan.. 
1936), No. 1, p. 21. 

DISCUSSION 

MR. TASKER: I should like to bring to this Committee the appreciation of the 
Sound Directors of Hollywood, as expressed at their meeting some ten days ago 
for the work that has been initiated by the Sound Committee, which, as Mr. Evans 
said, has very definitely stimulated a great deal of interest in solving this problem. 

We must realize that some of the differences in sound quality between the 
various studios are caused by the fact that our review rooms have substantially 
different characteristics. After hearing four or five outside pictures in one of 
our review rooms one of my men recently said, "Why is it that all the other 
studios have such bad sound and ours so good?" 

Although we must always expect differences of opinion, it is obvious that many 
of the differences that he had noted must have resulted from adjusting our re- 
cording characteristic to match our reviewing room characteristic, and certainly 
our product would not appear so favorably in another review room. As a 
result of this condition we are trying to get together on the reviewing room 
characteristic, which is the yardstick we apply to our respective products. When 
we have chosen some one yardstick, we shall then eliminate all but our wished- 
for differences in sound quality by making the corresponding revisions of our re- 
cording characteristics. 

MR. EVANS: The Sound Committee appreciates that this is a complicated 
problem, and is not only a matter of recording characteristics; the microphone 
characteristic is included, as well as the characteristics of the reproducing equip- 
ment, amplifiers, and loud speakers in the theater, and the acoustics of the theater. 

It was necessary to start somewhere on the problem, and apparently the 
Sound Committee chose a good point in getting the recording characteristics 
down on paper. 



REPORT OF THE COMMITTEE ON NON-THEATRICAL 
EQUIPMENT* 

At the request of the Standards Committee, steps have been taken 
to determine the preference of the industry with regard to the square- 
round or square-square combination of holes in reels used for 16-mm. 
film. The investigation followed as a result of publication 1 of the 
recommendations of the Deutsche Normenausschuss fur Kinotech- 
nik, which follow : 

(1) Spools having square holes on both sides are adaptable to all apparatus; 
whereas, the other spools do not fit on apparatus having square mandrels. Most 
German equipment has square mandrels. 

(<?) Spools made according to the American system are based upon operating 
the reproducing apparatus from the right, viewed in the direction of the light- 
beam. However, of late some apparatus operating from the left has been pro- 
duced; on such equipment only spools having square holes on both sides can be 
used, in order to be interchangeable on both left- and right-hand apparatus. 

(5) As there are two different standards for 16-mm. sound-film, distinguished 
by the position of the sound-track, it may be necessary in some cases to reverse the 
spool, if the film to be projected is of the standard other than the one for which the 
apparatus had been designed. This possibility is offered only by a spool having 
square holes on both sides.** 

A majority of the members of the Committee support the square- 
round combination, because it prevents rewinding or placing the film 
incorrectly upon the projector. Another point that seems to have 
been overlooked in the German publication is that an extremely 
great number of spools having the square-round combination are 
in every-day use throughout the world, and can not be employed 
upon projectors adapted to the square-square combination. If the 
spindles of the projectors are made square-round to accommodate 
reels of both types, a square-square combination will allow the reel 
to wobble, due to the inability to seat itself properly upon the spindle. 

Although it may appear that greater weight is on the side of the 

* Received October 15, 1936; presented at the Fall, 1936, Meeting at Roches- 
ter, N. Y. 

** Since this report was written, the SMPE standard fixing the position of the 
16-mm. sound-track has been approved by the International Standards Associa- 
tion, and thus has become a world standard. 
26 



REPORT OF NON-THEATRICAL COMMITTEE 27 

square-round combination, there is sufficient acceptance of the square- 
square recommendation to indicate that we may be forced to con- 
sider two alternatives, possibly making the square-round combina- 
tion the recommended standard and the square-square as an accept- 
able alternative. It may perhaps, be questioned whether a double 
standard would carry any serious objection; perhaps in the interests 
of international agreement a double standard might provide a neces- 
sary compromise. However, it is in order to point out that those 
who favor the square-square combination imply at least that that is 
not all that is desired, as their statements indicate that the film can be 
wound upon the reel the wrong way. Attention has been called to 
the patent situation ; one company holds patents on both the square- 
round and the square-square combinations here and abroad. Licenses 
on equitable basis could undoubtedly be arranged irrespective of 
what standard is finally employed. 

REEL CAPACITIES AND DIAMETERS 

There seems to be a considerable body of opinion that the Society 
should standardize reel capacities : for projector reels the sizes should 
be 400, 800, 1200, and 1600 feet, and possibly 2000 feet. This recom- 
mendation does not regard the 100-ft. return spools sent from the 
laboratory as requiring standardization. However, the 100-ft. 
spool might be discussed separately with the film manufacturers, who 
are primarily interested. 

The following hub and overall diameters are recommended for 
standardization : 

Capacity Diameter Hub Diameter 

(Feet) (Inches) (Inches) 

400 7 PA 

800 lOVs 4 7 /s 

1200 12V< 4Vs 

1600 13/< 4 7 /t 

2000 15 5 

PROJECTOR SPROCKETS 

The Committee has recently been asked for suggestions as to the 
desirability or possibility of standardizing projector sprockets. 
Although the Committee has not yet gone very far into the subject, 
it is well to emphasize the fact that such standardization presents 
great difficulties. For instance, the Bell & Howell Company employs 



28 REPORT OF NON-THEATRICAL COMMITTEE [J. S. M. P. E. 

two sprockets having slightly different pitch diameters. Other 
manufacturers employ only one sprocket. A detailed report on the 
subject will be offered in due course. 

EDUCATIONAL 

The use of motion pictures, particularly 16-mm. sound pictures, 
in the educational field continues to grow apace. At the recent 
N.E.A. Convention at Portland, this phase of educational activity 
received major attention ; and at a subsequent meeting at Hollywood, 
the newly formed Hollywood Educational Motion Picture Forum 
concentrated its entire efforts upon considering the applications of 
motion pictures to educational problems. The applications of mo- 
tion pictures have reached the point at which the regular theatrical 
producers can recognize the possibility of widening the outlets for 
their products or of supplementing their regular productions with 
education material. 

The U. S. Department of Education has just published the new 
National Visual Education Directory? which is a very complete 
survey of motion picture projectors and other visual aids in use 
throughout the country. However, even since the book has been 
placed upon the press, progress has been so rapid that its figures are 
rapidly becoming obsolete. 

BIOLOGICAL STANDARDS 

The following has been extracted from a report of the Committee on 
Standards for Motion Pictures of Biological Material, of the Biological 
Photographic Association : 3 

(1) Size. For general use, the 16-mm. size will usually be found preferable 
because of the lower cost and the more generally available projection facilities. 
Modern 16-mm. projectors give adequate size and brilliancy. 

For photographing, the choice between 35-mm. and 16-mm. film depends upon 
the subject and upon the finances available. The 16-mm. film has proved ade- 
quate for many biological subjects. It is recommended that 16-mm. prints of 
all subjects related to the biological sciences be available. 

(2) Color- Films. Color-films are preferable to black-and-white pictures when 
the colors of the objects photographed are important features. As the technical 
difficulties are reduced, more use should be made of color processes. Judicious 
use of stained or toned film or toned positives relieves the monotony of black-and- 
white, but any tendency toward "nature faking" is to be avoided. 

(3) Sound- Films. Sound-films are of value for special programs and for his- 
torical purposes, to bring to the audience the personalities of great persons. They 
are valuable for recording sound phenomena such as speech, animal sounds, diag- 



Jan., 1937] REPORT OF NON-THEATKICAL COMMITTEE 29 

nostic sounds, etc., when the original subjects can not be exhibited. Sound-films 
otherwise will probably have little use in college and university teaching because 
of their inflexibility. They can not be adapted readily to a given lecture or demon- 
stration. Cutting and re-editing film with sound-tracks is a difficult and costly 
process. A course is successful to the extent that it is attuned to the time, place, 
and students; and a sound-film for one purpose may not be satisfactory for an- 
other purpose. This objection may be met in part by running the sound-film 
silently when the sound is inappropriate or not required, provided the projector 
is so built as not to injure the film. As sound-films are projected at a speed of 24 
frames per second, more film must be used for the same interval of time. How- 
ever, the pictorial material in both a silent and a sound-film may be about the 
same because the titles in the silent film increase the length of showing when 16 
frames are projected per second. Sound on disks is another solution of the flexi- 
bility problem, as its use is adapted to both film speeds. 

(4) Silent Films will probably be used more frequently in class teaching except 
as noted in (5), above. 

(5) General Criteria of Excellence: 

(A) Photographic Technic. 

I. Sufficient illumination and freedom from flicker. (The latter 
is especially objectionable in time-lapse and animation 
technics.) 
II. Freedom from field jerks. 

III. Good placing and proper size of point of interest. 

IV. Uniform backgrounds (especially when films of several similar 

subjects are to be combined, as is often the case in surgical 
films). 

V. Correct composition and freedom from crowding. (The 
final image size upon the screen must be considered when 
the film is taken.) 

VI. Adequate use of long-shots and close-ups, to show orienta- 
tion and detail. (Dissolves are a useful means of shifting 
from the normal to the abnormal, and vice versa.) 
VII. Proper color value, definition and contrast, and freedom 

from glare. 

VIII. Proper use of accessories and related objects to produce a 
coordinated picture with respect to parts and to the entire 
film. 

IX. Efficient use of special technics (i. e., cinemicrography, 
dummy action, animation, accelerated or retarded motion, 
use of x-rays, stills, dissolves, etc.) 

X. Freedom from spots, scratches, etc. (Meticulous care and 
vigilance are necessary to keep all equipment clean and in 
good operating condition.) 

(B) Presentation (including editing). 

I. Consistent classification and transition from the general to the 

particular. 

II. Suitable spacing of striking phases to form climaxes. 



30 REPORT OF NON-THEATRICAL COMMITTEE [J. S. M. P. E. 

III. Adequate, properly arranged analytical and explanatory 

matter, adapted to the audience for which the film is 
planned. 

IV. Good continuity. (This may require the use of two cam- 

eras; e. g., during a surgical operation.) 
V. Absence of material unnecessary for good presentation. 
VI. Self-explanatory titles easily read and less than 15 or 20 
seconds long. Except for special instances */2 second per 
word is good. White block letters upon a black field are 
very effective, and the letters must be large enough to be 
readable at all audience distances. For long titles the 
moving script method is preferable. 

VII. Scenes long enough to justify the title, but not so long as to 
be tedious. (When necessary and possible, short shots 
should be repeated. Dissolves are useful when it is neces- 
sary to shorten a scene; e. g., the first suture may be shown 
at the end of a surgical operation, dissolving into the last 
suture to avoid unnecessary repetition.) 

(C) When films are planned for educational purposes the titles should: 

I. Be adequate to name the film, to show the location of it in 
space and time, and to explain the script from which it was 
made. 

II. Not be inaccurate, or so detailed that the film can not be 
adapted to different situations by means of running com- 
ment by the lecturer. 

III. Be too few rather than too many, when the film is to be 
used at the college or university level of instruction. 

IV. Give the proper generic and specific name to prevent in- 
evitable confusion from common names, which usually 
change with locality. 

(D) Cutting and recutting should be continued until the film can not be 

improved further by such means. 

(6) Subject Matter. Advantage should be taken of the motion picture technic 
to record what can not be shown better or more easily by other methods. A 
good motion picture must be more than a series of still subjects. The film should 
conform to professional ethics, and not contain undue commercial or personal ad- 
vertising. The material should be representative, rather than special or abnormal, 
when the film is intended for general use. The following lists suggested material: 
(.4) Life of organisms in their natural environment or showing their 

adaptability to unusual conditions. 

(5) Analysis of living processes by means of speeding or by showing at 
normal speeds phenomena not clearly seen; e. g., (a) animal and 
plant motions and behavior; (6) developmental processes of both 
plants and animals; (c) progressive reactions, as in parasitism, 
disease, etc. 

(C) Demonstrations, before an audience, of processes occurring on too 
small a scale for presentation or for which there is only one suit- 
able point of view. 



Jan., 1937] REPORT OF NON-THEATRICAL COMMITTEE 31 

(D) For the demonstration of a process, apparatus, and the proper tech- 
nics of investigation or presentation. 

(7) Intrinsic Value of the Finished Film. The real valueof a biological film 
depends upon how well it succeeds in achieving in the best way the purpose for 
which it was made, which can be decided only by audiences competent to judge its 
merits and effectiveness. Some films will have wide appeal, whereas others must 
be restricted to small groups of specialists. With increased dissemination of 
knowledge, the appeal of a really good film will increase, because what is spe- 
cialized knowledge today becomes general knowledge tomorrow. Consequently, 
a film should be planned, when possible, so that it may make the greatest contri- 
bution to biology and medicine. Sometimes long technical films may be cut and 
re-edited into shorter versions of interest to more general audiences. Certainly 
recognition will come to the photographers responsible for the production of films 
of exceptional merit. 

The Committee on Non-Theatrical Equipment will appreciate any 
comments or suggestions that may be offered to the mutual ad- 
vantage of both Societies for completing an efficient standardization 
along these lines. 

R. F. MITCHELL, Chairman 

D. P. BEAN E. C. FRITTS J. H. KURLANDER 

F. E. CARLSON H. GRIFFIN E. Ross 

W. B. COOK R. C. HOLSLAG A. SHAPIRO 

H. A. DEVRY A. F. VICTOR 

REFERENCES 

1 Die Kinotechnik (June 20, 1936), p. 197. 

s National Visual Education Directory, American Council on Education, Wash- 
ington, D. C. 

* "Report of the Committee on Standards for Motion Pictures of Biological 
Material," /. Biol. Phol. Assoc., IV (June, 1936), No. 4, p. 215. 

DISCUSSION 

MR. PALMER: Considering all the various kinds of film and apparatus it 
certainly would be a great help to have square holes on each side of the reel. 

MR. MITCHELL: In the report submitted to the Standards Committee several 
years ago we pointed out that the Bell & Howell Company was putting on the 
market reels having square holes. So many complaints were received from 
people who put the film on the projector the wrong way that we were forced to 
conclude that the disadvantages more than offset the advantages. 

MR. KELLOGG: How much would be added to the cost of a reel to add a little 
disk that could be slid from one position to another, changing the shape of the 
hole from square to octagonal, so that if one accidentally had the film wound the 
wrong way, he would not have to rewind twice? 

MR. FOSTER: The cost of an adapter inside the barrel, for making the hole 
either round or square, would be so small as to be negligible. 



REPORT OF THE STUDIO LIGHTING COMMITTEE* 

Several years ago the Studio Lighting Committee felt that it could 
serve the Society as well as the production branch of the industry to 
the best advantage by presenting a fairly complete compilation of 
studio illuminants, equipment, and lighting methods, with the thought 
that the material could be made available in the form of a handbook 
for cameramen and electricians. It was the plan to present the ma- 
terial in the form of four reports. The first was to consider the vari- 
ous illuminants available for studio photography, their characteris- 
tics, sizes, etc. The second was to discuss the equipment requirements 
as well as the many types available. The third report was to present 
data as to various sources of power, wiring systems, and control de- 
vices; and the fourth report was to discuss lighting practices, numbers 
of units necessary, methods of obtaining special effects, etc. 

Three of the reports, on illuminants, 1 equipment, 2 and sources of 
power, 3 have already appeared in the JOURNAL; but owing to the 
length of time that followed the publication of the third report, suc- 
ceeding reports have brought up to date the material already collected 
on illuminants and equipment. The fourth and final phase of the 
program, studio lighting methods, is discussed in this report. 

For purposes of study, motion picture lighting may be broadly 
classified into general and modelling. General illumination refers to 
relatively flat lighting of fairly uniform value, usually covering an 
entire set. In black-and-white photography it ranges from 200 to 
400 foot-candles. Its purpose is to provide sufficient light to make all 
parts of the set photographically "visible" and to provide a foundation 
upon which to build the modelling lighting. It serves to illuminate 
the shadows. Its counterpart in every-day life is the "skylight," or 
light dispersed by the atmosphere and reflected from the clouds. 

The lighting units that usually provide the general light are the 
twin broadside, the rifle, the scoop, and the multiple-lamp overhead 
unit, as well as many small auxiliary lamps placed behind posts, 
doorways, trees, etc. One arrangement often employed is to place 

* Received October 12, 1936; presented at the Fall, 1936, Meeting at Rochester, 

N.Y. 

32 



REPORT OF STUDIO LIGHTING COMMITTEE 33 

high-wattage lamps, such as the 200-, 300-, and 500-watt short tubu- 
lar projection types, in table and floor lamps as well as in wall brackets 
on the set, to supplement the general illumination and give the im- 
pression that the luminaires are lighted. On large sets, representing 
ballrooms, hotel lobbies, or outdoor scenes (being photographed in- 
doors), the usual general lighting equipment is not able to provide 
sufficient intensities at the distances involved, and resort is frequently 
made to the modelling units operated at their wider beam divergen- 
ces. Of course, scenes are frequently shown in which there appears 
to be no general lighting, the photography being done almost en- 
tirely with the modelling sources. This, however, is done to create 
certain special effects called for by the story. 

TABLE I 

Modelling Units in Current Use 

36-inch Reflector Spot 150-amp. arc 

10-kw. mazda lamp 
24-inch Reflector Spot 150-amp. arc 

5-kw. mazda lamp 

18-inch Reflector Spot 2-kw. mazda lamp 

80-amp. arc 
120-amp. arc 
150-amp. arc 

Lens Spots 10-kw. lamp 

5-kw. lamps 
2-kw. lamps 

Baby Spots 1000-w. lamp 

500-w. lamps 
"Lupe" Lights 1000-w. lamps 

Modelling lighting is, as the name implies, the part of the lighting 
that serves to create highlights and contrasts, and when properly 
used can serve as an excellent means of overcoming the two- 
dimensional limitation of the screen by producing a suggestion of 
depth. Its analogy in every-day life is sunlight. An additional 
function of modelling lighting is to place a much higher level of il- 
lumination upon the "star" or upon limited groups of actors, and thus 
direct attention to that part of the scene. Modelling levels of 
illumination are generally from two to four times as high as that of 
the general lighting, although in some instances the ratio may be as 
great as eight when necessary to produce desired effects. It is in 
handling the modelling lighting that the cameraman shows his ability. 



34 REPORT OF STUDIO LIGHTING COMMITTEE [J. S. M. P. E. 

A fairly common practice in many studios is for the "gaffer" 
or head electrician of the company to install and arrange the general 
lighting equipment, and then to have available an adequate supply 
of modelling equipment. The cameraman then devotes practically 
his entire efforts to placing and adjusting the latter. The customary 
modelling units are as shown in Table I. 

These units are all characterized by their ability to produce shafts 
of light having beam divergences of 12 to 20 degrees and sometimes 
less. Beyond angles of 20 degrees the illumination begins to ap- 
proach that of general lighting. By confining the light output to the 
relatively small angles stated, relatively high intensities can be pro- 
duced over the limited areas required of this kind of equipment. 

To produce the illusion of depth it is common practice to direct 
shafts of light downward from the rear of the set, producing strong 
highlights around the heads and across the shoulders of the actors 
and making them stand out from the background. This is frequently 
referred to as "back-lighting." In close-ups, the elements of general 
and modelling lighting are usually retained although the number of 
equipments is fewer. 

The customary procedure in studio lighting is to place the bulk of 
the modelling equipment, as well as a considerable part of the general 
equipment, upon parallel supports, suspended from the overhead 
structure of the stage. The scoops and domes or overhead strip- 
lights are hung directly from the overhead structures. The older 
practice was to build the parallels atop the walls of the set, but the 
use of unit flats for the walls has made necessary other means of 
supporting the parallels. 

When the camera position is approximately fixed, at least for a 
given series of shots, part of the general lighting units are arranged 
in a row extending in both directions from either side of the camera. 
A few modelling units are included, particularly those necessary for 
building up the illumination upon the star's face. There is, however, 
an increasing tendency toward greater numbers of "follow-up" or 
dolly shots, in which the camera, upon an easily movable carriage, 
follows the action at close range. For this type of shot the lighting 
units are often mounted directly upon the camera blimp, which ar- 
rangement has the advantage of maintaining balance of lighting at 
all times. 

It is well recognized that any chart that may attempt to indicate the 
numbers of the various kinds of lighting equipment for sets of several 



Jan., 1937] REPORT OF STUDIO LIGHTING COMMITTEE 



35 



kinds can afford only a rough approximation to the quantity of equip- 
ment that may be needed. The lighting needs of different cameramen 
vary greatly. The stories being filmed differ vastly in the quantity 
of light required. 

The very general rule is to mount upon the parallels as much equip- 
ment as space allows. The use of more compact equipment has, 
because of this rule, led to placing more equipment into position, 
although it may not all be used at one time. One might be tempted 
to remark, after seeing a Hollywood stage rigged out, that the car- 
penters who build the parallels determine the quantity of lighting 
equipment to be used. 

TABLE II 

Lighting Equipment for Motion Picture Photography 
(Black-and-White} 





Close Up 


Small Room 
(20 x 20 ft.) 


Medium Size 
(50 x 50 ft.) 


Large Size 
(Larger than 
50 ft.) 


Type of Set 


Min. Max. 


Min. Max. 


Min. Max. 


Min. Max. 


General Lighting 










Broadsides 


2 2 


4 8 


10 20 


20 40 


Overhead (Scoops) 




2 3 


5 10 


12 30 


Overhead Strips or Domes 


1 1 


2 3 


2 4 


8 12 


Modelling Lighting 










24- or 36-inch Reflector 




1 2 


2 8 


4 16 


18-inch Reflector 


1 2 


4 6 


6 12 


10 20 


Lens Spots (large) 


1 


1 


2 8 


4 16 


Lens Spots (medium) 


1 2 


2 6 


4 8 


12 24 


Lens Spots (small) 


1 1 


2 4 


2 4 


2 8 





from room to room, has made it necessary to place much of the light- 
ing equipment upon the parallels overhead in order to allow adequate 
clearances on the floor. It has further necessitated doubling or even 
tripling the lighting equipment for a given production unit, since all 
the rooms must be lighted simultaneously. 

Lighting for color motion picture photography does not differ in 
essential elements from lighting for black-and-white photography. 
The two outstanding differences are the much greater quantity of 
light required because of the filter loss and the necessity of splitting 
the light-beam as it passes the lens, and the importance of having 
light of the correct color quality. Since the color aids materially in 



36 REPORT OF STUDIO LIGHTING COMMITTEE [J. S. M. P. E. 

distinguishing depth, there is no necessity of creating extreme con- 
trasts, and flatter lighting may be employed. However motion 
picture lighting of the past years has been characterized by strong 
highlights and extreme contrasts, and many of the present color 
pictures retain this form of lighting. 

TABLE III 
Equipment Covered by Table II 



Broadsides Rifle 



Side Arcs 
Overhead Scoops Rifles 

Arc Scoops 
Overhead Domes or Strips 

24- or 36-inch Reflector Spots 

18-inch Reflector Spots 
Lens Spot (large) 
Lens Spot (medium) 

Lens Spot (small) 



1000- or 1500- w. lamps 
Twin 40-amp. white-flame arcs 
1000- or 1500-w. lamps 
Twin 40-amp. white-flame arcs 
1000- or 1500-w. lamps 
(usually five lamps) 
5000- or 10,000-w. lamps 
150-amp. arc 
2000-w. lamps 
120- and 150-amp. arcs 
2000- or 5000-w. lamp 
50-amp. arc 
100- and 500-w. lamps 



For the Technicolor process the general illumination level is about 
750 foot-candles and the highlight level 1000 to 1200 foot-candles. 
The photographing and processing are based upon light having sub- 
stantially 100 units of red, 100 green, and 100 blue, so as to permit 
color photography out of doors. For interiors, the white-flame arc 
has proved to be a fairly good duplication of daylight and has been 
used for general lighting. For modelling, the high-intensity arc, 
because of its abundance of blue-violet radiation, has been employed, 
although some filtering with amber gelatin is necessary to bring the 
blue-violet radiation into line with the red and green. 4 Incandescent 
lamps have been successfully employed, by using, in conjunction 
with blue filters, 6 high-efficiency types having a greatly increased 
blue output. 

In spite of the widespread adoption of faster camera lenses and the 
development of faster photographic emulsions, the tendency is to- 
ward more light, either through increasing the number of higher- 
wattage lamps, or by using more efficient lighting units such as the 
new Fresnel lens spots. 6 An inquiry was recently made as to what use 
was being made of the increased exposure resulting from these de- 
velopments, and the only satisfactory answer that could be obtained 






Jan., 1937] REPORT OF STUDIO LIGHTING COMMITTEE 37 

was that the film processing departments of several studios were 
taking advantage of it in respect to the greater latitude of develop- 
ment it afforded. In no instance was it found that the cameraman 
was employing this greater quantity of light and increased film sen- 
sitivity to stop down his lenses and thus attain improved depth of 
focus and better picture quality. 

New Equipment. Following the introduction of a new 2000-watt 
Fresnel lens spot, 6 one of Hollywood's well known equipment manu- 
facturers has made available a similar spotlamp in the 5000-watt 
size, and a still larger size has been developed employing a 150-am- 
pere high-intensity arc. Incandescent lamp manufacturers have 
recently placed upon the market a group of high-wattage lamps of 
the studio type, especially adapted, when used with a suitable filter, 
for Technicolor photography. They are designated CP as distin- 
guished from lamps used for black-and-white photography, which 
carry the marking MP. 

R. E. FARNHAM, Chairman 
W. C. KUNZMANN V. E. MILLER E. C. RICHARDSON 

J. H. KURLANDER G. F. RACKBTT F. WALLER 

REFERENCES 

1 Report of the Studio Lighting Committee, /. Soc. Mot. Pict. Eng., XVII 
(Oct., 1931), No. 4, p. 645. 

* Ibid., XVIII (May, 1932), No. 5, p. 666. 

8 Ibid., XXV (Nov., 1935), No. 5, p. 432. 

4 HANDLEY, C.W.: "Lighting for Technicolor Motion Pictures," /. Soc. Mot. 
Pict. Eng., XXV (Nov., 1935), No. 5, p. 423. 

6 FARNHAM, R. E.: "Recent Developments in the Use of Mazda Lamps for 
Color Motion Picture Photography," /. Soc. Mot. Pict. Eng., XXIV, (June, 1935), 
No. 6, p. 487. 

6 RICHARDSON, E. C.: "A Wide-Range Studio Spotlamp for Use with 
2000-Watt Filament Globes," /. Soc. Mot. Pict. Eng., XXVI (Jan., 1936), 
No. 1, p. 95. 

DISCUSSION 

MR. CRABTREE: Are any lamps available with the voltage control at the 
lamp, so that the cameraman can control the illumination locally without chang- 
ing the position of the lamp? 

MR. FARNHAM: The Vitachrome unit, used on the West Coast, has a small 
rheostat on the back, and quite a point is made of the light intensity control. 
The lamp consists of a reflector unit and a matte diffusing surface. 

MR. PALMER: Is any color photography being done using a combination of 
arc and incandescent lighting? 



38 REPORT OF STUDIO LIGHTING COMMITTEE 

MR. FARNHAM: At the present time difficulty is being experienced in getting 
correct balance. Incandescent lamps can be used successfully for Technicolor 
photography provided the light is unmixed. At the present time the incandes- 
cents seem to be a little deficient in the blue. We are engaged in the problem of 
correcting the difference so that the two kinds can be mixed. 

MR. RICHARDSON: Since the fast emulsions have been offered to the industry 
there has been a tendency to decrease the general front lighting a great deal, 
the reason being that there is considerable reflection from walls and floors that 
apparently provides part of the general lighting required. In fact, many sets 
have comparatively few of the floor lamps that used to be so popular. 

However, to meet the front-lighting problem, it is very common practice to 
mount two rather powerful lamps upon high standards at each side of the 
camera. In fact, camera bridges have been made structures under which 
the camera could be operated and on top of which could be placed lamps 
and operators for controlling the key lighting. 

The first Technicolor productions were made with a very considerable line-up 
of front lighting on each angle in proper balance with the camera, following the 
general lighting practice of three years ago. But now, in producing The Garden 
of Allah, for instance, the tendency was to utilize key lighting. Generally two 
150- or 120-ampere spots were used on each side of the camera. 

One thing that has made key lighting technic possible is that the new spot equip- 
ment has beam divergences so much greater than the old spot equipment, and yet 
it provides more satisfactory lighting than was possible with the general lighting 
arrangement of two lines of lamps on each side of the camera, generally diffused 
down to obtain the balance desired by the cameraman. 

MR. CRABTREE: Is the heat from incandescent lamps any longer a problem, 
since the sensitivity of the films has been increased? 

MR. FARNHAM: In the case of black-and-white pictures, we hear practically 
nothing of it any longer. For Technicolor lighting, with the increased level of 
illumination, heat is beginning to loom up as a problem, although the correcting, 
which is done by a filter blue in character and having some infrared absorption, 
tends to limit the heat somewhat. 



A THIRD-DIMENSIONAL EFFECT IN ANIMATED CARTOONS* 

J. E. BURKS** 

Summary. A description of the Fleischer Process by which animated drawings 
may be photographed in stop action, while superimposed upon a moving three-di- 
mensional background, and at the same time maintain the proper perspective for all 
positions. 

Since the process described here was developed at the Fleischer 
Studios, makers of animated cartoons, this paper will touch only upon 
the application of the process to cartoon production, which, however, 
is by no means the limit of its possibilities. 

Cartoon action was first produced by photographing pen and ink 
drawings on white paper. The method permitted only a limited use 
of backgrounds, if any, and the interest was naturally restricted 
solely to the action. Later, with the introduction of the celluloid 
process, the use of backgrounds rendered in all the tones required 
for well finished pictures first became possible. A single carefully 
executed background could be used for a complete scene, while the 
animated drawings traced in ink and filled in with flat opaque tones 
on transparent celluloid sheets carried forward the action by being 
photographed in sequence on top of this background. 

The completed picture was artistically correct only so long as 
the background was held in the fixed position for which it was drawn, 
and the perspective of the picture was necessarily determined by 
the camera angle. 

Now in cartoon presentation, panorama action, or, more correctly, 
following action, is highly essential in a large number of scenes. With 
flat backgrounds the following action is accomplished by sliding the 
drawing beneath the celluloid sheets at such intervals as to suit the 
action. In such case the background is rendered upon a long strip 
of paper. By this process, however, the perspective is "frozen"; 
that is to say, all converging lines move along at the same fixed angle, 

* Received August 22, 1936; presented at the Fall, 1936, Meeting at Rochester, 
N. Y. 
** Fleischer Studios, New York, N. Y. 

39 



40 



J. E. BURKS 



[J. S. M. p. E. 



T) 



and all distant points move through the field at the same speed as 
those in the foreground. 

In reality, in straight-line following action all points in a landscape 
or interior view appear to revolve about a point on the horizon at 
the center of vision. Actually the point is not on the visible horizon, 
but at the center of the field of vision at the distance of infinity. The 
process described here has been designed with particular attention 
toward simulating this apparent revolution of the converging lines 
in moving perspective. 

A truly three-dimensional background is used. A plastic miniature 

setting is photographed on a re- 
volving stage, so arranged that 
the center of revolution is the 
theoretical vanishing point of the 
set construction and is located in 
line with the center of the pic- 
ture plane and the camera lens. 
Fig. 1 is a plan view of the ar- 
rangement. Fig. 2 is the side 
elevation. A BCD is the stage 
platform, which swings about its 
axis at E. By experimentation 
a radius of five feet has been 
chosen as suitable for use with a 
lens of 50-mm. focal length. A 
cast aluminum platen, FG, is 
equipped with a double glass 
window and registration pegs 
upon which are placed the animated drawings of celluloid, the draw- 
ings being held flat between the two pieces of glass. The miniature 
background is constructed in an inclined plane extending from below 
the field at the edge of the stage to the desired horizon height on the 
circle HI, and is so arranged that the path upon which the animated 
figures are to appear to move is in line with the bases of the figures and 
the lens. Thus, when the perspective of the figures matches that of the 
setting, and the shadows are correctly drawn, the illusion that the 
action really occurs upon the setting is created. The various planes 
are brought into focus by using a small lens stop and a long exposure. 
Since the vanishing point of this arrangement has been moved from 
its natural position at infinity to a distance of about seven feet from the 




FIG. 1. Plan view of arrangement, 
showing camera, platen, and stage. 



Jan., 1937] THIRD-DlMENSIONAL EFFECT IN CARTONS 



41 



lens, all the objects in the scene must be built and arranged in forced 
perspective. That is, all receding lines that are perpendicular to the 
picture plane must converge to the center of rotation of the stage. 
A setting viewed from any but the proper angle is a queer sight to 
behold so far as perspective is concerned. 

The circle HI, which is the depth limit for all settings, is so selected 
that its diameter is somewhat larger than the limits of the camera 
field at that point. This permits a depth of about thirty-six inches 
for the horizon line of landscapes, which is ample to represent the 
distances of three to five miles usually occurring between the eye and 
the visible horizon. However, the relative speed at this circle is 
too fast to represent the passing action of very distant objects such 




FIG. 2. Side view of Fig. 1. 

as mountain peaks, which may be seen above the horizon as far 
away as twenty or thirty miles. We have overcome this problem 
by providing a secondary revolving stage inside this circle, as shown 
by JKLM, in Fig. 1. This is so geared to the larger stage as to turn 
at a speed half as great, and upon it are placed the mountain peaks 
and volcanoes. Fig. 3 shows a set with the machine ready for photo- 
graphing. 

There are some recent refinements to the process. One is a curved 
track located between the camera and the picture plane along which 
small automatically operated stages can move. The speed is so 
regulated that objects carried through the field appear to be a part 
of the set, but in reality are passing between the camera and the 
animated action. 

Another refinement is a stage that rotates vertically as well as 



42 



J. E. BURKS 



horizontally. This permits us to represent following action that 
rises into the air, vertically, diagonally, at any angle, or along a 
curved path. The additional axis is horizontal, and just back of 
and parallel to the horizon line AB. Thus, as the camera rises in 
height the foreground falls away, and the horizon line remains fixed. 
It has not been our purpose to duplicate nature or simulate reality, 
by the use of this process with animated cartoons. A cartoon is a 
drawing having much freedom of style, and so should be its back- 




FIG. 3. View showing set and machine ready for operation. 

ground. But any drawing, even a cartoon, to be correct artistically, 
must always be done in the proper perspective. So we build these 
backgrounds and paint them in the style of the cartoon, but the ma- 
chine keeps the perspective in order. The result is that so long as 
the stage is in motion about its fixed artificial vanishing point, the 
observer will be led to believe that he is looking into unlimited dis- 
tance instead of across only a few feet. The illusion is created by 
the difference of speed between the depth planes as each passes 
the observer, each plane moving in proper proportion to its distance 
from the revolving point. 

(A two-color sound picture, Musical Memories, was shown, in which appeared 
several scenes photographed by this process together with other scenes in which flat 
backgrounds were used. More recent releases use a three-color process.) 



MERCURY ARCS OF INCREASED BRIGHTNESS AND 
EFFICIENCY* 



L. J. BUTTOLPH** 

Summary. The characteristics of the high-pressure 85-watt mercury arc of in- 
terest to motion picture engineers are discussed, with special reference to brightness, 
dimensions, efficiency, utilization, and spectral quality. 

The low brightness of the Cooper Hewitt mercury arc, while an 
asset in industrial illumination, has kept the lamp from possible ap- 
plications in which high brightness, and consequent small source 
area are essential for use with reflectors and refractors, and where 
space is at a premium. The quartz mercury arc had a brightness 
of 500 to 1000 candles per square-inch, compared with the 15 candles 
per square-inch of the Cooper Hewitt lamp, and permitted some con- 
trol by large reflectors but almost none by condensers. This bright- 
ness has still been low compared with the 10,000 foot-candles per 
square-inch possible with incandescent lamps and the 100,000 
characteristic of the crater of the carbon arc, so little serious thought 
has been given to this mercury arc for projection or for long-range 
floodlighting work. 

About the only claim of the mercury arc to distinction as a photo- 
graphic light-source was the efficiency resulting from a high ultra- 
violet spectral content coincidentally with the high ultraviolet 
sensitivity of the earlier photographic emulsions. With the intro- 
duction of panchromatic film, the lack of red and blue-green in the 
mercury spectrum became a serious fault not subject to correction 
by filters. While the Cooper Hewitt lamp was always, in itself, 
essentially a direct-current device with a simple auxiliary, the recti- 
fier auxiliary for alternating current was rather bulky and heavy 
for its wattage rating. 

The recently developed so-called super-high-pressure mercury arcs 
embody sufficiently radical changes in brightness, spectral quality, 

* Received October 5, 1936; presented at the Fall, 1936, Meeting at Rochester, 
N.Y. 
** General Electric Vapor Lamp Co., Hoboken, N. J. 

43 



44 



L. J. BUTTOLPH 



[J. S. M. P. E. 



and auxiliary devices to open up some possibilities of studio or 
laboratory application. By designing a quartz mercury arc to operate 
at mercury vapor pressures of 20 to 30 atmospheres instead of at 1 
atmosphere, as did the older high-pressure arcs, a 
brightness of the order of 5000 candles per square- 
inch is attained in air-cooled lamps. By operating 
water-cooled arcs at higher pressure, brightness of 
100,000 to 250,000 candles per square-inch has been 
attained during rather short lamp lives. Of the 
possibilities ranging in rating from 50 to 10,000 
watts, only one unit has thus far been standardized 
for manufacture in the United States. 

The 85-watt, type H-3, mercury lamp (Fig. 1) 
may be thought of as a small version of the type 
H-l 400-watt and type H-2 250-watt mercury 
lamps 1 standardized during the past few years, 
except that it is not made for limited-pressure con- 
stant arc voltage operation, and so has a volt- 
ampere characteristic similar to that of the ordinary 
quartz mercury arc. It can and may be made at 
any time to possess the limited mercury charac- 
teristic of the H-l and H-2 types. Arc-heated, 
oxide-coated electrodes are carried on tungsten 
lead-in wires through a graded joint to a quartz- 
glass arc tube whose design may vary as indicated in Fig. 2. The 
function of the perforated diaphragms forming chambers at each 
end is to keep sputtered electrode material out of the arc tube. The 
diaphragms are not otherwise essential to the proper operation of 
the arc. A low pressure of argon provides the glow discharge 
necessary for the initial heating of the electrodes. 

As with the H-l and H-2 arcs, the H-3 is operated by a reactive 
transformer providing 440 volts for starting and 250 volts at the arc 
terminals at a normal arc current of 0.4 ampere. It is rated at 35 
initial lumens per watt in the arc, and for a life of 500 hours. The 
quartz tube of the arc proper is enclosed in an outer insulating bulb 
of glass that limits the short-wave radiation by absorption. Neither 
the shape nor the material of this bulb is critical, but a tubular bulb 
of hard glass is standard for the lamp at present, the dimensions being 
about those shown in Fig. 3. The glass limits the spectrum to about 
320 millimicrons . Through the visible and the near-ultraviolet range 



FIG. 1. Type 
H-3, 85-watt, 
mercury arc, 
base-up type 
(half size). 



Jan., 1937] 



MERCURY ARCS 



45 



the spectral distribution is similar to those of other high-pressure 
mercury arcs except for the unusual intensity of the 365-millimicron 
lines, which are as relatively strong as in the case of the standard 
Uviarcs. 

Although the dimensions of the arc tube are as shown in Fig. 2, 
the discharge is of the constricted type, giving a higher maximal 
brightness than the dimensions would indicate in calculation. As a 
light-source the constricted discharge is comparable with that of a 
ribbon filament incandescent lamp, and has the same advantages 



A B 

FIG. 2 Quartz-glass arc of H-3 lamp 
(full size): (A) original Philips type; (5) 
G. E. V. L. Co. and W. L. Co. type; (C) 
possible alternative without diaphragms. 

and limitations for use with optical systems. Aside from the effect 
of the size of the source upon the size of the optical system required 
to utilize it with practical efficiency, there is another relationship 
of similar significance, having to do with the effect of the length of a 
line-source upon the variation of intensity with distance. The 
theoretical relation is shown graphically in Fig. 4 in terms of foot- 
candles emitted by the 50-inch Cooper Hewitt, and by the 3 /4-mch H-3 
lamp, per hundred watts of total input. Even after adjustment is made 
for the difference of efficiency, it will be found that the smaller size 
of the source of itself permits obtaining a higher illumination. The 
limit, in each case, to the illumination attainable is the brightness of 






46 L, J. BUTTOLPH [J. S. M. P. E. 

the source as it approaches contact with the surface illuminated. 
The relation applies also when expressed, for example, in microwatts 
per cm. 2 as a function of distance, but foot-candles were used in Fig. 4 
because of their easy translation into actual illumination problems. 

While this arc is of the oxide-coated electrode type designed for 
a-c. operation, it may be operated on direct current by providing 
500 volts or more on a high induced voltage for starting and a series 
ballast resistance unit of about 600 ohms. The arc has a practically 
constant-voltage dynamic characteristic and so an approximate 





FIG. 3. Complete type H-3 lamp in two pos- 
sible bulbs (half size). 

sine wave current. Since the light output follows approximately 
the a-c. arc current, its intensity is variable, and although the flicker 
is not noticeable directly, it is such as to produce stroboscopic effects 
upon moving objects and may be a limitation in photography or 
projection where motion is involved. As a further result of these 
characteristics, the arc current and light might be modulated to some 
extent by applying a variable supply voltage, subject to the limita- 
tion that the minimal current must not be so low as to extinguish 
the arc, because the arc can not be restarted until it has cooled for 
several minutes. 

The relative energy distribution in the spectrum of the lamp is 
not closely comparable with that of any of the other standard mercury 



Jan., 1937] 



MERCURY ARCS 



47 



arcs, as shown graphically in Fig. 5, which is based upon unpublished 
data from B. T. Barnes of the G. E. Lamp Development Laboratory, 
and a recent paper by Elenbaas, 2 and by the spectrograms of Fig. 6, 




a 4 6 8 K3 

INCHES FROM SOURCE 



40 60 SO 100 



FIG. 4. The effect of source dimensions upon 
the variation of radiation intensity at vari- 
ous distances. 



which do not, however, do justice to the red end of the spectra. The 
most notable difference is in the red and in the 365 line, whose in- 
tensity is equalled only in the quartz Uviarc. The graphic comparison 
in Fig. 5 is based upon an equal over-all energy input of 100 watts to 
each lamp, and so gives directly the relative efficiencies. Some of 






48 



L. J. BUTTOLPH 



[J. S. M. P. E. 



these data, as calculated from the Barnes data, are given in 
Table I. 



TABLE I 



Microwatts per Sq. Cm. at 1 Meter, per 100 Watts Over-All Input, Based upon Reac- 
tive Ballasting of A-c. Units 



Wavelength 


Source 




C.H. 


H-l 


H-2 


H-3 


Uviarc 


7-600 




2.5 


1.4 


6.1 


0.3 


578 


5.9 


28.2 


18.3 


19.2 


17.8 


546 


18.1 


29.1 


19.2 


33.5 


16.1 


54-490 




1.7 


0.9 


4.3 




49-440 






0.4 


4.8 




435 


18.5 


20.4 


14.8 


27.1 


14.2 


405 


13.1 


9.9 


7.9 


14.5 


8.7 


365 


6.5 


9.1 


7.5 


25.7 


25.1 



Overall Watts 



450 



450 



300 



100 



410 




560 600 640 680 
54-6 578 60O-7OO 



FIG. 5. A comparison of the relative energy distribution in the principal 
lines of typical mercury arcs. 



Apart from applications of the H-3 lamp to conventional types of 
projection enlargers and to high-fidelity ultraviolet sound recording 
of a nature beyond the scope of this paper, no unique application to 
motion picture engineering is suggested. It is believed, however, 
that the high intensity of the 36>5-millimicron line and the high 
brightness of the source may permit application of the lamp to certain 
of the more highly specialized lighting problems of the industry. 




FIG. 6. Spectrogram of the arc. 

REFERENCES 

1 ST. Louis, J. A. : "Characteristics of 400-Watt and 250-Watt Type H Mer- 
cury Lamps," Trans. I. E. S., 31 (June, 1936), No. 6, p. 583. 

BUTTOLPH, L. J.: "High-Intensity Mercury and Sodium Arc Lamps," /. 
Soc. Mot. Pict. Eng., XXIV (Feb., 1935), No. 2, p. 110. 

1 ELENBAAS, W.: "Die Intensitatsverteilung und die Gesamtsrahlung der 
Super-Hochdruch-Quecksilberentladung," Physica (Holland), 3 (Aug., 1936), 
No. 8, p. 859. 



DISCUSSION 

MR. TASKER: When the lamp is operated on direct current, at 500 volts and 
with 600 ohms ballast resistance, is the life reduced to less than what it would 
be on alternating current, or is it about the same? 

MR. BUTTOLPH: Probably reduced; the reason being, of course, that the 
cathodic functioning of the electrode is the more severe, and on direct current 
it is continually on one electrode. 

MR. WOLF: I am rather disappointed to learn that this high-intensity source 
has not found wider application in projection and photography. What are the 
real reasons for the inability to use it? Is it the spectral characteristic, or the 
variation in intensity? 

MR. BUTTOLPH: I did not mean to suggest that it could not be used for pro- 
jection. That is one of the possibilities. Even this little source might be used 
in some forms of projection. The Philips Company anticipates extensive applica- 
tion of the water-cooled type of arc for projection. 

MR. TASKER: Do they expect to attain greater screen illumination than we 
can attain with other types of arcs, or merely an economy in electrical power? 

MR. BUTTOLPH: I should say both, but the water-cooled type of arc seems to 
open up interesting possibilities of increasing the intrinsic brilliancy. Apparently, 
brilliancies can be attained quite comparable to those attained with the higher- 
intensity carbon arcs. 

MR. BRENKERT: Did you refer to the possibility of the alternating current 
producing a flicker in the projected image? 



50 L. J. BUTTOLPH [J. S. M. P. E. 

MR. BUTTOLPH: Yes. The a-c. arc causes flicker at the rate of 120 cycles per 
second. 

M,R. BRENKERT: Has the flicker been found to be objectionable? 

MR. BUTTOLPH: I am not in a position to discuss the technicalities of projec- 
tion, but I understand that by synchronizing the shutter to the flicker of the light- 
source the results are very satisfactory. 

MR. BRENKERT: But without such an arrangement, is the flicker discernible 
on the screen? 

MR. BUTTOLPH: I do not know. 

MR. WEBER: How does the heat from the arc compare with that from the 
incandescent lamp for a given candle-power? Is it a cooler source of light than 
what we are using now? 

MR. BUTTOLPH: Yes, by perhaps one-half. That is an advantage inherent 
in all arcs. 

MR. de JONG: When using direct current, the idea is to change the polarity 
every time the arc is used. 

The Philips Company is developing a system of projection without any shutter 
by causing the dark periods of the lamp to occur at the right instants. The 
system is not yet ready to be disclosed, but that shows how things are progressing. 
As far as the light is concerned, I can not present figures on our tests and trials, 
although we have created some very high intensities on the screen. 

MR. TASKER : By "very high intensities" do you mean higher than are attainable 
or comparable with the highest otherwise attainable? 

MR. de JONG: Comparable with the highest we had in mind. 

MR. DILLEMUTH: What is the voltage of the Philips lamp? 

MR. de JONG : Between four and six hundred, across the arc. 

MR. JOY: Must the diameter of the light -source always be small compared 
with the length? 

MR. BUTTOLPH: In general, yes, to attain these high brilliancies and high 
efficiencies. Since with arcs of this type we utilize only a limited source area, the 
arc can be shortened. There is a definite limit to the extent to which the arc can 
be broadened and yet operate efficiently and supply high intrinsic brilliancies. 

MR. JOY: Would it depend upon the voltage of the arc? 

MR. BUTTOLPH: No, it depends upon the thermal instability of the gas through 
which the discharge occurs. If we were to build the lamp in the form of a bulb, 
and then attempt to operate it at these high pressures, a constricted discharge would 
wander throughout the bulbular space. It would not take the shortest line be- 
tween the two electrodes. In other words, the wall of the arc physically defines 
the position of the discharge. At high pressures, even in a large bulb, the dis- 
charge would be of this very ribbon-like character. 

MR. SPONABLE: When operating a water-cooled arc is it customary to obtain 
the water from the regular water supply, or do you use a closed system? 

MR . de JONG : We use a closed water-cooling system, and use the same quantity 
of water all the time. Water from the city mains is not harmful over very long 
periods of time. The water absorption is very low. 

MR. KELLOGG: Speaking of using the arc for projection, is it not necessary 
to supplement it with some other form of light, or to use a filter in order that the 
color distribution will be reasonably satisfactory for the eye? 



Jan., 1937] MERCURY ARCS 51 

MR. BUTTOLPH: I assume that the arc would be limited to some extent to 
black-and-white projection. There is considerably more red, and the ratio between 
the yellow and the green is changed so that the light is definitely yellower, not 
quite daylight, and yet not incandescent light. 

MR. KELLOGG: If it is desirable to mix the light with that of tungsten, it is 
quite conceivable that you could produce an optical image of a tungsten filament 
practically coincident with the arc itself. If that were done, suppose the tungsten 
filament were directly behind the arc; would the arc absorb a great share of the 
light of the tungsten filament, or is the absorption discontinuous, and effective 
over only limited ranges of wavelength, so that the tungsten filament would 
contribute a good deal of light? 

MR. BUTTOLPH. I believe that most of the tungsten light would go through, 
so to speak. There might be some absorption in the portion of the spectrum where 
the mercury lines lie, but it would not be serious. Why not do the reverse: why 
not put the image of the mercury arc hi tfre plane of a tungsten filament grid? 

MR. WOLF: Can Mr. de Jong tell us something about the extremely high- 
intensity light-sources being used on airplane landing fields in Holland, and 
whether or not they are applicable to photography? I saw one at Eindhoven, 
and I never saw a source that could equal it in brilliancy. 

MF. de JONG: We have some lights that are used for landing fields, which 
can be concentrated very nicely, and will light up almost the entire field. I 
can not tell exactly how the system works. 

As to photography, very good pictures can be made with the lamp, especially 
by combining floodlights and spotlights with the vapor lamps. The lamp is now 
at the point of coming into practical use, and I do not believe that I can tell very 
much more about it. 

MR. JOY : Why is it not possible to start the lamp immediately when it happens 
to go out? 

MR. BUTTOLPH : Starting an arc of this kind is a matter of ionizing the mercury 
vapor. We have no way of doing so at a pressure of more than a fraction of an 
atmosphere, even with very high voltages. What all the reasons are, I am unable 
to say. The deionization time is fairly long, and will carry over nicely even on 
25 cycles. But once the arc is entirely out, it is difficult to reestablish it. 

MR. FARNHAM : Answering Mr. Kellogg's question, in some experimental work 
on the lamp, when we tried to project inccjidescent light through the arc screen, 
the lens characteristics of the quartz tube interfered somewhat. Possibly 
more can be accomplished by using a better shape of tube, perhaps hi the form of 
a cylindrical lens. 

We have had some success in starting the arc immediately after it has gone out, 
by using the high-frequency discharges of an ultraviolet ray machine, which 
start the arc immediately. 

MR. BUTTOLPH : That is our regular laboratory practice. We have been unable 
to apply it practically to lamps operating at full temperature. 

MR. KELLOGG: When we were discussing the possibility of superimposing 
the arc image upon a tungsten filament, it did not occur to me that the hot 
tungsten might be used as a mirror to reflect the arc, in addition to producing its 
own radiation. Such a plan, however, seems to me to present some practical 
difficulties, whereas the other should be quite easy. 



STANDARDIZATION OF MOTION PICTURE MAKE-UP* 

M. FACTOR** 

Summary. The use of make-up is traced briefly from its earliest application in 
the theaters of ancient Greece to the present time, and the relation between make-up 
and the various forms of lighting from one period to another is alluded to. 

The paper continues with a detailed discussion of the relation between the charac- 
teristics of the present make-up components and the color-sensitivities of various types 
of film. Data are given for the various make-up materials in units of color according 
to the Lovibond tintometer glasses; photographic reflecting powers are given; and 
curves are presented showing the spectral transmittance curves of the Lovibond glasses 
to match various make-up materials. 

HISTORICAL 

The use of make-up dates back to the earliest historical times. 
We know that it was used in all the ancient civilizations. During 
the Golden Age of Greece, the leisure classes used the various make-up 
mediums such as are used today. They knew of powder, foundation 
creams, lip pomade, mascaro, and nail staining. In Rome, during 
the time of Christ, the application of make-up became an art. The 
hair was bleached as well as dyed, and certain portions of the body 
were tinted; for example, the knee-caps and the bottoms of the feet 
were colored and the toes were tinted. The make-up powder known 
as Cyprius, came from the island of Cyprus. 

In the theaters of both Greece and Rome, make-up had a position 
of lesser importance. The players, instead, depended upon masks 
for the various characterizations. However, during the revels of 
Dionysius, certain of the revelers, who were to mourn for dead per- 
sons, stained their bodies in appropriate symbolic colors. From these 
revels sprang the Greek tragic plays and the origin of the theater. 
In the Grecian theater, make-up was not used for character protrayal; 
but instead, the characters were indicated by the use of masks of 
various shapes and colors to symbolize the characters. In the be- 
ginning, by changing masks, one actor portrayed all the various parts. 

* Presented at the Spring, 1935, Meeting at Hollywood, Calif. 
** Max Factor, Inc., Hollywood, Calif. 

52 



STANDARDIZATION OF MAKE-UP 53 

Primitive masks were used in the ceremonial dances and theatricals 
of aboriginal peoples. In the Aleutian Islands, New Zealand, Siam, 
and other countries, the masks may still be seen. In the ceremonials, 
in addition to the masks, the bodies were often covered with grease 
and earthy pigments. The American Indian, who did not use masks 
as a rule, painted his features for war dances. In certain ceremonies 
he used masks representing animals. 

The use of masks continued in practically all forms of dramatic 
entertainment until about 1600, the time of Queen Elizabeth, when 
they were discontinued and pigmented powders substituted for them. 

Between 1500 and 1600 the art of make-up was indeed crude. 
The actors at that time very carelessly made themselves up to repre- 
sent Harlequin, Punchinello, and other characterizations of the period. 
They often used white chalk dusted over their bodies, and wigs for 
their heads. Make-up was of secondary importance. 

The theaters were crudely lighted by oil-burning lamps, tallow 
candles, and iron baskets with resinous woods. Because of the sub- 
dued lighting, make-up in light tones was used, and from that time 
on actors handed down their make-up secrets from generation to 
generation. j* ; 

After 1800, stage lighting changed from crude oil illumination to 
the better, more efficient gas lighting. Through the passing years 
the art of make-up and theater illumination depended upon each 
other. Prior to 1850, dry colors were applied directly upon the 
natural skin or dusted upon a grease foundation. 

About 1880, when the electric light was invented, a new impetus 
was given the theater and a new demand placed upon make-up. It 
was about this time that stick grease paint made its first appearance, 
as an invention of a German actor touring the United States. 

By 1890 the making of stage make-up materials came into such 
wide use that their manufacture attained a commercial scope. About 
1900, when the first motion pictures were in practical use, make-up, 
borrowing the technic from the stage, came into use when the first 
studios installed electric lights. 

The first motion picture make-up was of a pink color, due to the 
fact that the photographic emulsions were of the orthochromatic 
variety. However, since orthochromatic emulsion is not sensitive 
to the yellow-red of the spectrum, the faces of the players with pink 
make-up photographed in an artificial looking tone. 

It may be recalled; that the very early motion pictures were cen- 






54 M. FACTOR I J. S. M. P. E. 

tered around more or less standardized types of characters, among 
which were the villain, the hero, the heroine, and perhaps the father. 
The villain usually did not apply a coat of make-up that would have 
improved his appearance a little too much: a dark blemished skin 
was more to be desired. The villain usually wore a long mustachio, 
and heavily painted eyebrows. The hero, on the contrary, applied 
make-up in quantities. His lips were usually elaborately painted, 
and his face was as white as the heroine's, which was, indeed, white. 
The heroine affected long curls and bore much make-up. 

The secondary players in those early dramas depended upon a coat 
of the pinkish grease paint and penciled eyebrows. 

In the photographic processes, make-up is used to add certain pig- 
mentations to those of the human face so that the photographic re- 
production will coincide with the visual; that is, for correct and nat- 
ural photographic reproduction, the face must monochromatically 
reproduce in a range such as the eye normally visualizes. 

The eye sees yellow, more readily than any other color, with a rela- 
tive proportion of more red than blue. The photographic emulsion, 
on the other hand, does not record in the same manner. As is known, 
the panchromatic photographic emulsion has a high sensitivity in the 
blue, with a proportionately lesser sensitivity in the red. 
. The characteristics of the Mazda illumination, on the contrary, as 
used on the motion picture sets are relatively low in the blue and very 
high in the red end of the spectrum. The characteristics of photo- 
graphic emulsions that are particularly sensitive to the blue give rise 
to the condition of a too dark color rendering in the red end of the 
spectrum and too light in the blue. 

. The human face without make-up does not photograph smoothly 
because of the irregular distribution of the pigmentation that gives 
color to the skin. This difficulty is remedied in still photography by 
suitable retouching. Make-up in motion picture photography serves 
to correct the condition by presenting to the camera a well balanced 
and uniform distribution of color tones that reproduce upon the 
screen in monochrome an effective interpretation of what the eye 
would normally see. The purpose of make-up, then, is to add to the 
face sufficient blue coloration in proportion to red in order that the 
photographic tonal rendition will be such as the eye sees in real life, 
and to prevent excessive absorption of light by the face. 

The psychological problem must also be, considered. A combina- 
tion of the correct blue- violet and yellow-orange with sufficient blue 



Jan., 1937] STANDARDIZATION OF MAKE-UP 55 

in a range that would be desirable for a correct panchromatic re- 
cording would result in a brownish make-up. On the motion picture 
set before the camera, this brownish make-up would distinctly affect 
the psychological responses of the players ; for example, if two play- 
ers were enacting an emotional scene and had to look at each other in 
unnatural colors, a certain interference of response would result. 
Even though containing the desirable panchromatic colors in the 
make-up composition, the make-up must be sufficiently natural to 
overcome this psychological factor. 

This is accomplished by adding to the composition of the make-up 
a medium that has a comparatively high reflection in the yellow- 
orange range (about 6000 A). In compounding the make-up with 

TABLE I 

Make-Up Composition According to the Lovibond Tintometer Glasses 

Make-Up Material Units of Color 

Minus Green Minus Blue Minus Red 

No. 3 1 Panchromatic foundation 9.5 7 . 00 1 . 00 

No. 31 Panchromatic powder 4.70 3.00 0.06 

No. 21 Panchromatic foundation 3.00 0.95 0.00 

No. 21 Panchromatic liner 10.00 1.70 4.21 

No. 5 Moist rouge 10.00 1.70 4.80 

No. 9 Moist rouge 16.00 10.00 7.00 

No. Studio Special rouge 14.00 7.00 6.00 



this problem in mind, the dominant hue of the make-up material is 
held as near 6000 A as possible; at the same time sufficient pigments 
of the valuable photographic colors are added to assure correct photo- 
graphic rendition while still maintaining a semblance of natural visual 
coloration of the skin. 

This problem is solved to an extent by balancing the colors in the 
make-up; that is, a certain relative proportion between the yellow 
and blue in contrast to the red is maintained. To give body and to 
hold make-up in a natural skin color range, sufficient quantities of 
the red-yellow and non-color-selective material are added. 

The different densities and panchromatic values of make-up are 
controlled by thus balancing the three unit colors. The effect of 
these three colors in relatively equal concentrations results in a gray 
or black when applied to the face, as they act as subtractive colors, 
while a lesser color concentration of any one of the elements alters the 



56 M. FACTOR [J. S. M. P. E. 

hue in favor of the other two. For example, in the panchromatic 
moist rouges, series 5 to 9, the color concentration increases propor- 
tionately more in the yellow and red than in the blue (very little blue 
is needed to affect the make-up color materially). The increase in 
color results in a make-up that is darker both visually and photo- 
graphically. 

This proportionate increase may be seen in Table I. The data 
were obtained with the Lovibond tintometer and glasses,* for analy- 
tical studies by the subtractive method, and the researches and com- 
parisons were conducted by Capt. H. B. Haselden of the Henry E. 
Huntington Library, R. E. Farnham, G. A. Chambers, and W. E. 
Theisen in collaboration with the Max Factor staff. 

The Lovibond tintometer glasses were used in lieu of other color 
analyzing systems because of their practicability. The color ab- 
sorption of the glasses was analyzed and data presented by Gibson 
and Harris 1 of the U. S. Bureau of Standards. According to these 
authors, certain irregularities appeared in the graduation of the Lovi- 
bond glasses, although for purposes of the present study of make-up 
these errors are too small to be of sufficient importance. 

In making the analysis of the color compositions of the make-ups 
in terms of the Lovibond tintometer glasses, a Bausch & Lomb com- 
parison microscope was used. The usual calcium sulfate screen, 
which is employed as a standard white in the Lovibond tintometer, 
was placed in position so that light reflected from it filled one-half of 
the two-part comparison field. A sample of the make-up to be 
measured was likewise positioned so that light reflected from it filled 
the other part of the comparison field. The calcium sulfate screen 
and the sample were both illuminated by sunlight. In order to obtain 
a color balance in the comparison field, suitably selected Lovibond 
tintometer glasses were positioned so that the light reflected from the 
calcium sulfate screen was filtered thereby, thus modifying the color 
of the comparison field until it matched the part of the field illumi- 
nated by light reflected from the make-up sample. From a knowledge 
of the spectral absorption of the various Lovibond glasses found 
necessary to produce a match, it was possible to compute the spectral 
composition of the light in the comparison field. These spectral 
absorption data were taken from those published by Gibson arid 
Harris. 1 

'----.'* jf. W. Lovibond, Salisbury, England, 



Jan., 1937] 



STANDARDIZATION OF MAKE-UP 



57 



In Fig. 1 the dotted curves represent the spectral transmittance 
values for the 3.0 minus green and the 1.0 minus blue Lovibond tintom- 
eter glasses which were found necessary to match the No. 21 pan- 
chromatic foundation. By multiplying together the ordinates of the 
two curves, wavelength by wavelength, the solid curve is obtained, 
which gives the spectral transmittance of the combined glasses. 
This curve shows the spectral distribution of radiant energy that will 
produce a subjective match with the No. 21 panchromatic foundation. 
It should be kept in mind, however, that the spectral reflectance of 



100 - no. 21 Panchromatic Foundation 




4500 



5000 



5500 



6500 



7000 



FIG. 1. Spectral transmittance curves for 3.0 minus green and 1.0 
minus blue Lovibond tintometer glasses to match No. 21 panchromatic 
foundation. 



the No. 21 panchromatic foundation may not necessarily be the same 
as that indicated by the solid curve of Fig. 1. It is quite possible to 
find different spectral distributions of radiant energy that will pro- 
duce a subjective match with the distribution as indicated by the 
solid curve. This follows from the fact that the eye, in terms of 
which the color match was made, is not an analytical sensitive recep- 
tor. It is well known that two colors that appear identical to the 
eye may differ markedly in spectral composition. However, the 
solid curve of Fig. 1 is of considerable value as an. indication of the 
appearance of the No. 21 panchromatic foundation. For instance, 
the fact that the ordinates of this curve in the region between 5500 
and 7000 A are relatively high, as compared with the ordinates in the 






58 



M. FACTOR 



[J. S. M. P. E. 



region between 4000 and 5000 A, indicates that the No. 21 panchro- 
matic foundation is reflecting a predominance of the longer wave- 
lengths, and is absorbing rather strongly the shorter wavelengths 
of the visible spectrum to which panchromatic photographic ma- 
terials are relatively quite sensitive. 

In Fig. 2 is shown a similar analysis of the data relative to the No. 
31 panchromatic foundation. In this case three Lovibond tintometer 
glasses were required to produce the desired color match, and the 
dotted curves represent the spectral transmittance of the three glasses 



No. 31 Panchromatic Foundation 



U.O Minus Green x 




4000 



7000 



FIG. 2. Spectral transmittance curves for 3.0 minus green and 1.0 
minus blue Lovibond tintometer glasses to match No. 31 panchromatic 
foundation. 



used, while the solid curve is computed by multiplying the ordinates 
of all three of the components together, wavelength by wavelength. 
Here again the solid curve gives the spectral distribution of the light 
transmitted by the three glasses combined, and hence indicates a 
spectral distribution (but not necessarily the only one) that will 
match the No. 31 panchromatic foundation. 

In Fig. 3 the same analysis is shown for the studio special rouge. 
The fact that the ordinal values of the solid curves of Figs. 2 and 3 are 
appreciably less than those of the solid curve of Fig. 1 indicates that 
the No. 31 panchromatic foundation and the studio special rouge have 
appreciably lower total reflectances than the No. 21 panchromatic 
foundation. This conclusion is verified by the data shown in Table II. 



Jan., 1937] STANDARDIZATION OF MAKB-UP 59 

In order to ascertain the photographic reflecting power of the make- 
up, L. A. Jones, of the Kodak Research Laboratories at Rochester, 
N. Y., furnished data obtained with the photographic reflectometer, 
described in 1932 2 in the JOURNAL. The Lovibond readings give a 
subjective evaluation of the color of the make-up materials as seen 
by the eye, while the photographic reflectometer values translate 
these colors into their relative monochromatic brightness as seen by 
the photographic emulsion. 

In measuring the photographic reflecting powers of the samples, the 



100 



Studio Special Roue* 



90 

1.0 Minus Red 




4000 4500 5000 5500 6000 6500 7000 
Wavelength 

FIG. 3. Spectral transmittance curves for 3.0 minus green and 1.0 
minus blue Lovibond tintometer glasses to match studio special rouge. 

illumination of the samples was derived from a tungsten lamp operat- 
ing at a color temperature of approximately 3400 K. This illumi- 
nation was incident at an angle of 45 degrees from the normal to the 
surface, and observation was along the normal to the surface. The 
photographic material, in terms of which the evaluation was made, 
is the Eastman supersensitive panchromatic motion picture negative. 

The rouge and foundation were used in layers sufficiently thick to 
be practically opaque. The powders were dusted over a thin film of 
the foundation, the application being quite generous so as to attain a 
fairly complete coverage with the powder. Table II shows the values 
obtained. 

Although the measurements were not made under what might be 
called practical studio conditions, the quality of the light was vir- 



60 M. FACTOR [J. S. M. p. E. 

tually identical to that of the Mazda illuminants used upon the set. 
The readings, therefore, represent the integrated actinic effect of the 
color with respect to the light-source and the color-sensitivity of the 
emulsion used for the determination. 

The need of a standard make-up was particularly emphasized with 
the introduction of the panchromatic emulsion. As the result of a 
series of experiments made at Warner Bros. Studio in Hollywood in 
February, 1928, the Max Factor panchromatic make-up was adopted. 
Subsequently to these experiments, the panchromatic film, the Max 
Factor panchromatic make-up colors, and Mazda illumination were 
introduced. 

TABLE II 

Photographic Reflecting Powers 

(Per Cent) 

No. 5 Moist rouge 390 light 14 . 2 

No. 9 Moist rouge 390 dark 7.6 

Moist Studio special rouge 5.2 

No. 21 Panchromatic foundation 61.9 

No. 31 Panchromatic foundation 23.3 

No. 21 Panchromatic powder 70.5 

No. 31 Panchromatic foundation powder 38.8 

No. 22 Panchromatic liner 6.1 

To assist with these problems the author set out to introduce a 
standardized series of make-up colors. At that time, because of lack 
of suitable colorimetric instruments and a knowledge by the industry 
of its needs in this respect, a trial-and-error method was used to de- 
velop a practicable motion picture make-up. The primary purpose 
was to balance the make-up colors correctly, and to recognize the 
relations between the motion picture mediums. Second, it was 
necessary to present a make-up material in a form that would be 
practicable. Because the old theatrical stick make-up was not easily 
applied and was unsanitary, the unguent form of make-up, with a 
base composed of vegetable oils, was introduced in a collapsible tube. 
Soon the old theatrical stick make-up was discarded in favor of the 
unguent panchromatic foundation color. 

The unguent base aided in the manufacturing processes since the 
pigments could be introduced in closer balance and could be added 
in finer and smoother forms. This makes possible a thinner 
application, permitting a freer use of the facial muscles and leading to 



Jan., 1937] STANDARDIZATION OF MAKE-UP 61 

better acting by the players. Because of the fineness of the pigment 
in the unguent panchromatic make-up, it is possible to apply a de- 
sired protective color coating adequately without an unduly heavy 
application of the make-up. 

By experiment it was found that a range of panchromatic colors, 
numbered from 21 to 31, would be sufficient for motion picture use. 
From the 21 pan foundation make-up, with no blue in its composition, 
the photographic absorption increases proportionately up to the 31 
pan. Referring to Table I, this increase from the 21 to the 31 is in 
the photosensitive colors; that is, in the blue end of the spectrum. In 
this range the photographic reproduction of the progressive numbers 
becomes proportionately darker. As the colors darken their yellow 
and blue contents increase, with relatively less increase in the red 
pigmentation. 

The panchromatic powders ranging from No. 21 to 31 correspond 
in color pigmentation to the unguent foundation of the same numbers. 
Visually, because of the weaker total light absorption of all the spec- 
tral colors, they appear relatively lighter than the corresponding pan 
foundation. The powder serves the purpose in make-up practice of 
eliminating sheen and reflection, and of fixing the unguent foundation. 
It is desirable to use the powder of the same number as the foundation. 

The moist rouges in a paste-like unguent are designed chiefly for 
lip make-up. The red and blue are balanced sufficiently to attain 
the desired photographic tonal contrast with respect to the founda- 
tion colors. The studio special rouge, which is widely used in motion 
picture practice, visually appears redder than the series 5 to 9 rouges, 
due to the greater proportion of red and yellow in its composition. 

Correct application of make-up and subsequent correct balance of 
illumination and photographic exposure result in the desired photo- 
graphic image in monochrome. When this balance is correctly at- 
tained the features of the players are well defined, and a certain round- 
ness or perspective is maintained. Because the camera records in a 
two-dimensional plane, it is one of the problems of make-up so to 
control the photo-reproduction of the features of the players that their 
faces reproduced upon the screen have the appearance of roundness. 
The natural facial form is suggested by building contrasts of light and 
shadows, and the make-up composition is such as to give a correct 
photographic rendering in monochrome. 

Throughout the various cinematic moods and characteristics of 
the settings, there is a great variation of intensity of illumination 






62 M. FACTOR 

upon different parts of the set. For example, the face in strong, direct 
light reflects more yellow and orange, while the blue is, to an extent, 
absorbed. By contrast, in indirect illumination this condition is 
changed, and the reproduction has a tendency to be darker than 
natural. The results are different, too, in light containing an abun- 
dance of infrared, as is the case with certain of the studio lights and 
not so with other lights upon the same setting. The make-up pro- 
vides a protective coloration that is theoretically a balance for a vari- 
able lighting condition. 

It is the problem of make-up so to compose its colors and inherent 
characteristics that a wide photographic tonal range can be covered. 
That is, the player must be able to progress through light beams from 
various sources without too sharp contrasts. The made-up face must 
reflect relatively more light in the shadows than in the direct lighting. 

Make-up, besides furnishing the necessary protective coloration for 
the photographic processes, is also used to aid characterization. 
Angular faces may be softened by a cleverly balanced use of light and 
darker patches of make-up. Wide faces may be made to seem 
narrower by artificially adding shadows to the sides of the face. 
Certain features must be defined and outlined in keeping with the 
desired characterization. Many effects and results, which are be- 
side the point of this paper, are gained by an artistic application of 
make-up. Make-up is an artist's tool, and for clever characteriza- 
tions its application depends upon artistic skill. The make-up artist 
must have a sense of light and shade, facial contours, and the rela- 
tionship of the face to the desired characterization, all based upon the 
photographic reproduction in monochrome. 

The author is particularly indebted to Mr. E. W. Theisen and Dr. 
L. A. Jones for assistance in the preparation of this manuscript. 

REFERENCES 

1 GIBSON, K. S., AND HARRIS, F. K. : "The Lovibond Color System; a Spectro- 
photometric Analysis of the Lovibond Glasses," Scientific Paper No. 547 (Feb. 17, 
1927), p. 1, U. S. Bureau of Standards, Government Printing Office, Washington, 
D. C. 

8 JONES, L. A., AND MCFARLANE, J. W.: "The Precise Measurement of Filter 
Factors and Photographic Reflecting Powers," /. Soc. Mot. Pict. Eng., 
(Oct., 1932), No. 4, p. 364. 






A RECORD WORD-SPOTTING MECHANISM 
R. H. HEACOCK** 



Summary. The record word-spotting mechanism described here places the pick-up 
needle upon a predetermined spot on a phonograph record by pressing a remotely 
located release button after three reference readings, previously established by the 
trial and error method, have been properly set. The pick-up arm is held poised 
above the record by a direct electromagnetic pull upon the back end of the pick-up arm. 
When the electromagnet is de-energized, the pick-up falls due to the pull of gravity. 
The speed of fall may be controlled by means of an adjustable exhaust port on an air 
dashpot. No catches or latches are used in releasing the arm. A manually operated 
open-circuiting release button is in parallel with a second open-circuiting switch 
in the electromagnet circuit, the second switch being opened each revolution of the turn- 
table by a cam. To release the pick-up, the manually operated button is depressed, 
but the pick-up is not released until the second switch is cammed open by the turntable. 
In this way the device is indexed with relation to the radial position of the record so 
that not only may the correct groove be repeatedly selected, but the desired portion of 
the groove may be consistently repeated. The effect of eccentricity of the center hole 
of the record with relation to the recorded grooves is also eliminated. Variations in 
the size of the record hole are accommodated by means of a tapered centering pin. 
Each of the mechanical parts, with the exception of the cammed turntable switch, is 
rigidly located on the single pick-up arm unit. All necessary electrical parts for 
complete operation of the mechanism on 105- to 125-volt, 50- to 60-cycle alternating 
current are located on the under side of the motor board. 

HISTORICAL 

One of the main differences between an entertainment program of 
today and that of ten years ago is the naturalness of atmosphere of 
the presentation. This is true whether the source be the stage, 
talking picture, or broadcast studio. On the stage, sound reinforcing 
systems assure clear reception of every sound by each patron in the 
theater, no matter how far he may be from the footlights, and these 
systems also permit the easy introduction of certain background and 
supporting sounds which greatly add to the realism of the per- 
formance. Naturally, talking pictures and broadcast studios are 

* Received September 21, 1936; presented at the Fall, 1936, Meeting at 
Rochester, N. Y. 
** RCA Manufacturing Company, Inc., Camden, N. J. 

63 



64 R. H. HEACOCK [J. S. M. P. E. 

faced with broader sound effects requirements since many of the 
scenes portrayed are supposedly happening under unusual conditions, 
as in aeroplanes, automobiles, trains, or upon busy highways, where 
it is essential that background sounds provide the necessary "color" 
in the establishment of the idea of seeing and hearing these actual 
occurrences. We are all familiar with the many sounds that are 
later dubbed into the final sound-track. What leaves Hollywood 
with only an original dialog on the sound-track may have many 
different sounds dubbed in to indicate a change of scene from a 
busy street to a quiet lobby and then to a hilarious night club dance 
floor. In the broadcast studio, Captain Henry's "Showboat" with 
its churning paddle wheel and deep whistle, Fred Allen's cheering 
crowd as it enters Town Hall and the Town Hall News with its falling 
curtain, "lights out," and clicking projector add a very necessary 
touch of realism to the performances. 

Many of the simpler sounds are made by special gadgets. When 
sheet cellophane is lightly crumpled it gives the cheery crackle 
of the campfire. If crumpled more slowly with somewhat greater 
pressure, the sound becomes that of frying bacon and eggs. When a 
clothesbrush is drawn very, very slowly across a taut drumhead, 
a downpour of rain is reproduced, and eccentric swishing of the 
brush upon the drumhead produces the ocean's roar on the beach. 
Two plumber's vacuum cups, when beaten upon the demonstrator's 
chest, reproduce Paul Revere's horse, and when transferred to a 
table top the horse's pounding upon a wooden bridge is accurately 
reproduced. In addition to many of the simpler sounds, it is es- 
sential for the proper presentation of many programs that sounds 
that are much more complex be obtained from already existing 
phonograph records, e. g., a record is cut in immediately following an 
actor's cries of "Fire!" reproducing the roaring motors, screaming 
sirens, and yells of the excited crowds. 

In the past, sound effects devices have been constructed in an at- 
tempt to make the recorded sounds available for use in the regular 
program. In general, the limiting and most disappointing feature 
of such devices has been the spotting mechanism for placing the 
needle upon the record at a predetermined point. Units have been 
constructed with high-gain amplifiers and high-fidelity speakers 
so that it has been possible to control the volume satisfactorily, as 
well as the response, through the use of convenient mixer networks 
to mix and control the overall output level properly when actual 



Jan., 1937] WORD-SPOTTING MECHANISM 65 

sound is desired or by means of mixer network alone when it is 
desired to mix electrically. 

The spotting mechanism, however, has generally been a rather 
complicated and not entirely trustworthy device. Latches that are 
supposed to be released when, and only when, an associated electro- 
magnet is energized frequently engage the pick-up arm release so 
securely that the electromagnet can not disengage the latch, and 
instead of the expected sound, nothing but an empty period of 
silence, lasting until some quick-thinking person steps in to "cover 
up," follows the moment the sound effect device has been cued in. 
If the contact face of the latch has been altered to permit easier 
disengagement, frequently the arm will drop upon the record ahead 
of the time desired if a slight jolt should disturb the equipment. 

Since other spotting mechanisms used in the past have been released 
entirely independently of the radial position of the turntable with 
relation to the pick-up arm, it is evident that if a record is used, the 
center hole of which is eccentric with relation to the recording grooves, 
a variation in the point of contact between the needle and the record 
will be introduced, which is extremely serious and which may amount 
to several grooves, one way or the other, depending upon the radial 
position of the record with relation to the pick-up arm. 

DESCRIPTION 

The difficulties that have been outlined above have been carefully 
considered in the design of the spotting mechanism, and it is felt that 
each has been overcome. The record word-spotting mechanism is a 
device that will, with extreme accuracy, place the pick-up upon a 
predetermined spot on a record, so that the sounds recorded at this 
predetermined spot may be reproduced and mixed with the regular 
program. The mixing process may be accomplished acoustically or 
it may be accomplished electrically. Acoustical mixing involves the 
use of amplifiers and speakers, while electrical mixing makes use of a 
proper mixing network to feed the pick-up output into an external 
line. 

The pre-locating tone arm (Fig. 1) consists essentially of an up- 
right tone arm support in which is mounted the swivel arm member 
of the tone arm. A tilting member, at the outer end of which is 
affixed the turntable pick-up, is mounted in a horizontal swivel in 
the divided portion of the swivel arm. The tilting portion of the tone 
arm is held up, by means of an electromagnet, in such a manner that 



66 R. H. HEACOCK [J. S. M. P. E. 

as the tone arm is swung laterally the pick-up may be poised, at any 
point in its swing, over the record. An adjustable counterweight is 
provided on the tilting arm to regulate the pressure at the needle 

point. 

A release push-button (Fig. 1) is connected in parallel with another 
switch located on the motor board (Fig. 2). When the release push- 
button is depressed, and the latter switch opened by means of a 
fixed cam on the under side of the turntable, the magnetic circuit is 
opened, thus permitting the tilting arm to fall and the pick-up needle 



DRSHPOT 



COUNTER 
WEIGHT 



ELECTROMflGNET 




RELERSE 
BUTTON 

I 

f 
FIG. 1. Mechanism in operating position poised above record. 

to engage the record. Since the cam switch opens the circuit of the 
electromagnet always at the same radial point on the turntable, 
regardless of the time within the revolution that the release push- 
button is depressed, the pick-up needle always engages the record 
at a fixed radial angle from the switch-opening cam. 

A white spot on the edge of the sound effects record (Fig. 2) is 
indexed with relation to engraved figures upon the turntable rim. 
The record is centered and excess clearance eliminated by means of a 
tapered centering pin (Fig. 2). 

The speed of fall of the pick-up is determined by the adjustment of 
a small dashpot attached to the rear of the tilting arm (Fig. 1). In 
order to locate the radial position of the arm accurately with relation 



Jan., 1937] WORD-SPOTTING MECHANISM 67 

to the record, there is located on the top of the vertical pivoted portion 
of the tone arm and rigidly attached thereto a gear engaged by a 
pawl (Fig. 3). This pawl is supported by a pawl plate located im- 
mediately above the gear so that a notch in the plate will expose one, 
and only one, tooth of the gear at a time (Fig. 4). Each exposed tooth 
is of a different color, permitting indexing the tone arm for radial 
position. 

A tilting adjustment (Fig. 3) is provided upon the upright support 
of the tone arm so that when the pick-up is at rest (i. e., not playing 




FIG. 2. Turntable and associated parts. 

a record) it will rotate slowly toward the right in such a manner 
that the heel of the pawl will engage the spindle of a micrometer head 
(Fig. 4). This micrometer head is equipped with a dial gauge, giving 
an outer dial and an inner dial reading. It is to be noted that one 
revolution of the indicator of the outer dial is equivalent to the 
movement of the indicator from one number to the next upon the 
inner dial. Care must be exercised to see that the pawl is properly 
engaged with the correct tooth on the tone arm gear. 

The colored teeth on this gear provide the main sectors of index- 
ing, with the accurate setting in any one sector shown upon the dial 



68 



R. H. HEACOCK [J. S. M. P. E. 



gauge. Thus, both the color of the tooth and the reading of the dial 
gauge determine the radial position of the arm with relation to the 
record. Consequently, with the white dot on the record placed op- 
posite one of the numbers on the turntable rim, the proper setting with 
relation to the gear color and dial gauge reading will insure the accurate 
placement of the needle upon a predetermined spot on the record. 

It is to be noted that as the white dot on the record edge is moved 
clockwise or counter-clockwise, syllables or notes may be subtracted 
from or added to the original point of contact of the needle on the 
record. 



GERR 

PHWL LEVER 

TILTING SCREW 




FIG. 3. Pick-up arm adjustments. 

Since both motions of the tone arm mechanism are dependent for 
operation upon gravity alone, and no latches or catches are present, 
there will be no sticking or jamming of the mechanism. 

From the foregoing description of the device, it is evident that the 
precision with which the mechanism operates makes it possible to 
index any specific sound in a card file by means of the record number 
and three separate readings ; namely : 

CO The color of the gear tooth exposed by the notch in the pawl plate. 

(2) The inner dial and the outer dial readings of the micrometer dial gauge. 

(3) The location of the white dot on the record with respect to the engraved 
numbers on the turntable rim. 

In this way, the usual "hunting" for the desired sound is eliminated, 



Jan., 1937] WORD-SPOTTING MECHANISM 69 

and such sounds are made as fully available as the various volumes 
in a properly indexed library. 

OPERATION 

To reproduce a desired sound, the procedure should be as follows : 
The record is placed upon the turntable, and the centering pin is re- 
placed firmly in order to center the record and remove any excess 
clearance between the record and the pin. The tilting adjustment 
screw is set so that the arm rotates slowly toward the right when 



SINGLE CtflR 
TOOTH 



PHWL HEEL 




FIG. 4. Pick-up arm with gear light removed. 

held poised above the record. The micrometer head is set to indicate 
approximately zero on its dial indicator. The record is played in 
the normal manner, and when the desired sound is heard the phono- 
graph motor is immediately shut off. The left forearm is rested lightly 
on the tone arm in order to prevent dislodging the needle from the 
record groove. The pawl is then released by drawing the pawl lever 
away from the axis of the tone arm support with the index finger of 
the left hand. The pawl plate is rotated until the pawl heel is brought 
into contact with the spindle of the micrometer head. The pawl 
lever is then released so that it will engage the nearest gear tooth 
away from the spindle of the micrometer, and the pawl is locked into 
mesh with the gear by applying pressure on the pawl lever toward 



70 R. H. HEACOCK [J. S. M. P. E. 

the axis of the tone arm support. The micrometer head is next 
adjusted until its spindle contacts the pawl heel, and the spindle is 
backed off approximately thirty units on the outer scale of the dial 
indicator, to compensate for the coasting of the motor after power 
has been shut off. The pick-up is raised so that the electromagnet 
holds it poised above the record. The mechanism is then released by 
depressing the release button and it is noted whether the needle 
contacts the record ahead of or behind the desired sound. Then the 




FIG. 5. Sound effects phonograph unit for National Broadcasting 
Company. 



micrometer adjustment is moved in or out, as may be necessary, 
keeping in mind that approximately ten points upon the outer scale 
of the dial indicator is equivalent to a movement of one groove at 
the needle point. Continue in this manner by the trial and error 
method until the needle consistently engages the proper groove. 
Precise adjustment is then made by rotating the record with relation 
to the turntable until the pick-up needle falls upon the exact point 
of the record for the desired sound. A rotation of the record through 
approximately ninety degrees will add or subtract (depending upon 
the direction of rotation) a short word or a syllable. 



Jan., 1937] WORD-SPOTTING MECHANISM 71 

ADVANTAGES 

The device has the following advantages : 

(1) Simplicity: There are no locking catches or mechanical 
latches used in either the up and down motion of the pick-up arm or 
in the rotary motion of the arm. Gravity alone controls both the 
motions. Consequently, there is no possibility of sticking or jam- 
ming, and it is not necessary to make any very accurate adjustments 
of either of these motions. 

(2) Once the needle is engaged by the record groove, the arm is 
as free to move in any direction as a manually operated arm. There 
are no springs or other constant loads tending to force the needle 
in any direction. 

(3) Since the pawl heel is in engagement with the micrometer 
head until the needle is actually pulled away by the record groove, 
there is no possibility of inaccurate lateral wavering of the pick-up 
arm after release or before engagement with the record groove. 

(4) Since the tapered centering pin accurately centers the record 
by removing all "play" between the record and the pin, and since the 
rotary position of the record with relation to the turntable is fixed, 
it is possible to replace the record upon the turntable and bring the 
needle down upon the exact word or sound that may be desired by 
pre -setting the device. 

(5) Even though recorded grooves may be eccentric with relation 
to the center hole, the device will still function perfectly satisfactorily. 

(6) Since the beginning of fall of the pick-up arm is indexed by 
the record itself, it is possible to spot a fixed sound accurately very 
much more readily than with a device that is not indexed according 
to the rotary position of the turntable. 

(7) Each of the mechanical parts, with the single exception of 
the cammed turntable switch, is rigidly located upon the single pick-up 
arm unit. This makes it possible to move the arm and use it in com- 
bination with any turntable with satisfactory results, without locating 
the parts accurately with relation to each other on each new set-up. 

POSSIBLE USES 

Figs. 5 and 6 show a typical example of a de luxe application of 
the mechanism. Fig. 5 is the phonograph amplifier unit, and Fig. 6 
is one of the loud speaker units recently developed and manufac- 
tured for the National Broadcasting Company for producing sound 
effects in their various key studios throughout the country. Al- 



72 R. H. HE ACOCK 

though this record word-spotting mechanism was originally developed 
for the National Broadcasting Company for this purpose, it is evi- 
dent that the same release mechanism could be used in combination 
with existing turntable units, amplifiers, and speakers, so that the 
overall cost of the set-up could be readily controlled, depending 
upon the specific requirements of each installation. 

The overall equipment is extremely flexible in view of the fact that 
the single fully equipped arm may be set up on any motor board 




FIG. 6. Sound effects loud speaker unit for National 
Broadcasting Company. 

for use with any turntable as long as the turntable has been equipped 
with the proper release cam. 

Since the release button may be located remotely the device 
lends itself to stage operation, since the mechanism could be located 
in the wings and the stage director could personally release the pick-up 
to provide a realistic scream or other desired sound, which could be 
properly cued with relation to the rest of the performance. The 
device is also useful in dubbing sound from already recorded sources 
on either wax or film. It is satisfactory for use in automatic announc- 
ing systems or for any other purposes where certain definite sounds 
must be reproduced from sources already recorded on records. 






A DEVELOPING MACHINE FOR SENSITOMETRIC WORK* 
L. A. JONES, M. E. RUSSELL, AND H. R. BEACHAM** 

_ Summary. Sensitometric testing of photographic materials requires that the 
laboratory be able to obtain the same results, with a high degree of precision, for 
identical samples of material, although the individual tests may necessarily be made 
at widely different times. All factors tending to influence the results must be held con- 
stant over long periods of time. 

A developing machine is described designed for a laboratory in which a relatively 
large volume of sensitometric work must be done. It accommodates sixty strips 
positioned vertically on six metal racks, which can be lowered into the developer 
simultaneously and removed either simultaneously or individually, so that different 
development times may be given to different parts of the load. 

The developer circulation across the face of the exposed material is sufficiently 
rapid that further increase of agitation produces little if any increase in the rate of 
conversion of latent image into metallic silver. The circulation is of two sorts: (/) 
a relatively slow, uniform movement of developer in the vertical direction produced 
by a propeller that forces the developer down into a well external to the main tank, 
from the lower end of which it spreads out beneath a perforated false bottom in the 
tank and rises throughout the body of the tank, flowing back again into the top of the 
well; (2) a much more violent agitation produced by vertical paddles moving back 
and forth close to the exposed surfaces. Both agitating elements are driven by a 
synchronous motor, assuring the same rate of circulation at all times. The entire 
machine is jacketed by thermostatically controlled water at 65 =*= 0.1F. 

Results show that the circulation throughout the body of tht, tank is so nearly uni- 
form that they are not influenced by (a) whether the heavily exposed end of the sensito- 
metric strip is up or down, (b) the position of the strip within the tank, or (c) whether 
a complete or partial load of strips is developed at one time. Results indicate also 
that the agitation is sufficiently violent that the rate of conversion of latent image into 
metallic silver is at or near the maximum attainable. Uniformity and reproducibility 
of development is very markedly superior to that attainable with any type of hand- 
or machine-rocked tray with which the authors have had experience, and the use of 
the machine marks a very definite advance in the precision with which sensitometric 
values may be established. 

The process of measuring the characteristics of a photographic 
material by sensitometric methods may be divided, for convenience 
of discussion, into several distinct steps: (a) selecting the sample; 
(b) conditioning the sample prior to exposure ; (c) exposing the sample 
* Received October 9, 1936; presented at the Fall, 1936, Meeting at Roches- 
ter, N. Y. Communication No. 600 from the Kodak Research Laboratories. 
** Eastman Kodak Co., Rochester, N. Y. 

73 



74 JONES, RUSSELL, AND BEACHAM [J. S. M. p. E. 

in a sensitometer; (d) processing the exposed sample; (e) measuring 
the densities in the developed sample ; (/) plotting the data in graphic 
form; and (g) interpreting the results. 

In order to obtain consistent and reproducible values over long 
periods of time, every step in this process must be controlled to a 
high degree of precision. In the present communication, we are 
concerned chiefly with an instrument and a method for obtaining 
a high degree of reproducibility in processing the material which has 
been exposed in the sensitometer. 

The processing may also be separated for convenience of dis- 
cussion into several steps, namely : development, fixation, washing, 
and drying. Of these, development is by far of the greatest impor- 
tance. While it is true that variations in the washing, fixation, and 
drying of the developed sensitometric strip may produce some varia- 
tions in the density of the resultant silver deposits, these variations 
are relatively small, and it is quite easy to maintain the conditions 
of washing, fixing, and drying sufficiently constant so that no varia- 
tions can be attributed to these processes. On the other hand, it 
is extremely difficult to maintain the constancy of development with 
sufficient precision to eliminate errors arising from variations in the 
development process. For instance, the density of the resultant 
silver image depends upon the constitution, concentration, and 
temperature of the developing solution, and also upon the time of 
development, and very profoundly upon the degree of agitation of 
the developing solution. 

The present discussion is confined largely to methods of maintain- 
ing constancy of temperature, time, and agitation. 

During past years in most sensitometric laboratories the develop- 
ment of the exposed sensitometric strips has been carried out in hand- 
rocked or possibly machine-rocked trays. Where the number of 
strips to be processed is not particularly large and the work can be 
done by one operator, the hand-rocked tray method is fairly satis- 
factory. An individual apparently can maintain a rocking technic 
over relatively long periods of time that will give consistent results, 
but it seems to be very difficult for two or more operators to learn 
rocking technics that will give results agreeing with each other. 
Hence, if the volume of work is such that several operators are re- 
quired, the hand-rocked tray method leaves much to be desired. 

Many efforts have been made to develop mechanical means for 
rocking trays that will not give definite wave patterns and result in 



Jan., 1937] MACHINE FOR SENSITOMETRIC WORK 75 

local nonuniformity of development action. Some of these have been 
quite satisfactory but not entirely so. Here, again, a relatively large 
number of such devices is required to handle a large volume of work, 
since a single tray can not accommodate a large number of sensito- 
metric strips. On the whole, the mechanically rocked tray has not 
proved very satisfactory where a large volume of high-precision 
sensitometric work is to be done. 

Many attempts have been made to build developing machines for 
this purpose, and some very successful results have been obtained. 
We shall not attempt to discuss the literature at length, but a bibli- 
ography covering some of the more important efforts in this direction 
is appended to this communication. Suffice it to say that in most 
cases the design has been from a very specialized point of view, and 
the machine has not been of such a nature as to accommodate a large 
number of strips at one time ; nor have these designs been such as to 
be susceptible of redesign in order to accommodate larger numbers of 
strips. 

The enormous increase in the volume of sensitometric work done 
in this laboratory forced us about two years ago to consider the design 
and construction of a sensitometric developing machine, and in the 
following pages will be given a description of this instrument, to- 
gether with a discussion of the quality of the results attained there- 
with. 

In designing a developing machine to meet our requirements, 
it was considered desirable to obtain as closely as possible the follow- 
ing characteristics: 

(1) It should be precisely reproducible from physical specifications. 

(2) It should produce highly uniform development over that area of the ma- 
terial contributing to the results of the test. 

(5) The degree of agitation of the developer should be such that any variation 
would produce little or no variation in the resultant densities. 

(4) The capacity of the machine should be such as to permit processing a 
relatively large number of strips at each loading. 

(5) The quantity of developer per strip should be relatively small. 

(6) It should be possible to remove part of the strips at various times without 
disturbing the remainder. 

(7) The machine should be easy to operate in total darkness. 

(8) Changing the developer and cleaning the machine should require a 
minimum of time. 

(9) The design should permit the use of a convenient and accurate automatic 
temperature regulator to hold the temperature of the solution within the re- 
quired limits. 



76 



JONES, RUSSELL, AND BEACHAM [J. S. M. P. E. 



(10) The machine should be built of materials not affected in any way by the 
various developing solutions that must be used. 

(11) The structure should be rugged and simple so as to require a minimum of 
upkeep. 

In Fig. 1 is shown a cross-section through the developing machine 
installed in the water bath used for maintaining constant temperature. 
The developing solution is contained in a rectangular stainless steel 
tank about 20 inches long, 6 inches wide, and 16 inches deep. The 
sensitometric strips, which are 12 inches long and 25 mm. wide, are 




FIG. 1. Cross-section of the developing machine installed in the 
water bath. 

mounted upon racks and positioned vertically in the developing 
tank. Fig. 3 is a schematic view of the developer tank with the 
positions of the sixty strips indicated. 

At the rear of the rectangular tank in which the strips are developed 
is a cylindrical well about 3 inches in diameter. This is connected 
to the bottom and top of the developing tank so that the developer 
can be forced downward in the well by a propeller type of stirrer 
driven by a synchronous motor. The structure is such that the 
developer driven down in this cylindrical well is delivered uniformly 
to all parts of the bottom of the tank, entering under a perforated 
false bottom. It spreads out and rises uniformly through the cross- 



Jan., 1937] 



MACHINE FOR SENSITOMETRIC WORK 



77 



section of the tank, and, as it reaches the top, flows back into the 
open upper end of the cylinder. The mechanical structure is such 
that uniform upward velocity of the developer is maintained at 
every point in the tank in which the strips are being developed. In 
addition to this circulation of the developer, much more violent agita- 
tion is provided by a set of three vertical paddles which are moved 
back and forth at a fairly high velocity. 

As stated previously, the 12-inch sensitometric strips are mounted 
upon stainless steel racks. When the strips are prepared, two holes 



1J I i j i r 




FIG. 2. 



Schematic view of the driving mechanism, motor, and method of 
suspending the strips. 



are punched at each end. By means of these holes, the strips are 
mounted upon pins, one pair of which is carried by a coil spring so 
that the strips are maintained under uniform tension at all times. Ten 
strips are mounted upon each rack and provision is made for the 
placement of six racks in the tank at one time. 

In Fig. 1, the heavy vertical black lines represent the positions of 
the racks carrying strips, while the cross-hatched areas represent 
the paddles. It will be seen that a paddle is positioned between each 
pair, the emulsion side of the strips facing inward toward the paddle 
between each pair of racks. These paddles are driven back and forth 
by a synchronous motor operating through a worm and worm- 
wheel and a double alternating rack and pinion movement, as in- 



78 JONES, RUSSELL, AND BEACHAM [J. S. M. p. E. 

dicated in Fig. 2. The velocity of travel of the paddles is such that 
a complete cycle is made in approximately six seconds. 

A superstructure, also of stainless steel, is built onto the tank so 
that the film holders may be placed in the proper position before 
being lowered into the solution. Guides are built in the tank to hold 
the racks in place during development. These guides extend a dis- 
tance above the surface of the solution so that the operator is able 
to place the lower end of each rack in its guide while the rack is sus- 
pended above the level of the developing solution by means of two 
arms extending outward from the superstructure. When all is in 



POSITION NUMBERS IN DEVELOPING MACHINE 




J_ _J_ _3_ _*_ _S_ _fi_ _2_ 


_A- _a_ 


Jfl. 


_!L Ji_ Jl_ J- " JS. -IL Jl. 


js_ _a. 


. 


^L 1*. -|3_ 24_ J5_ J6_ 27_ 


12. 12. 


JO. 


_*L _ d3_ *i- .i*. -li- ^L 


48. 49_ 


50 


51 52 53 54 55 56 57 


56 59 


60 



FIG. 3. Schematic diagram indicating the positions of the sixty 
strips in the developer tank. 

readiness, the operator pushe.8 downward on the carriage from which 
are suspended the film racks. This carriage is counterbalanced by 
weights so that it stays in its upper position until forced downward 
by the operator. When the carriage is lowered and the strips are 
immersed in the solution, the stirring mechanism automatically 
starts. The two arms supporting the film racks may now be turned 
back and the carriage returned to its upper position, thus making 
it possible to remove at any time one or more of the racks without 
disturbing the others. 

When a rack is taken out, it is quickly rinsed in the acid stop-bath 
mounted immediately in front of the machine (Fig. 1). From there 
it is transferred directly to the fixing bath which is positioned con- 



Jan., 1937] 



MACHINE FOR SENSITOMETRIC WORK 



79 



veniently to the left of the developing machine. The fixing bath 
container is also water- jacketed by the constant-temperature bath 
so that all solutions to the action of which the strips are subjected 
are at the same temperatures. 

It is usually desirable when developing sensitometric strips to use a 
series of development times in order that the effect of time of develop- 
ment upon the sensitometric characteristics may be determined. It 
has been necessary, therefore, to remove a part of the strips at a series 
of different times, and, since most of this work is done in extremely 
dim illumination or even in total darkness, it is necessary to have 




FIG. 4. 



Photograph of the assembled developing machine, 
mechanism is shown at the extreme right. 



The timing 



some automatic timing device that will indicate to the operator when 
a given rack of strips should be removed. For this purpose a special 
timing device was constructed and mounted conveniently near the 
machine so that the operator receives both a visual and an audible 
signal. This timing mechanism is driven by a synchronous motor 
which is started automatically by the lowering of the carriage that 
supports the racks carrying the exposed strips. This synchronous 
motor, operating through a double worm reduction gear, drives a 
set of six disks so that they make one complete revolution in 20 
minutes. Upon the periphery of each of these disks is a lug that may 
be set at any desired position relative to the starting point. As the 
group of disks rotates after the starting of the motor, these lugs, by 



80 



JONES, RUSSELL, AND BEACHAM [J. S. M. P. E. 



closing electrical circuits, give the desired signal to the operator. 
The first signal is given by a buzzer which begins to operate a few 
seconds prior to the completion of the development time for which 
a particular lug is set; and in addition to this, a small electric lamp 
located behind a green translucent window, upon which is a number 
corresponding to the film rack to be removed, is energized. At the 
instant the buzzer stops, the indicated time has elapsed, and the 




FIG. 5. 



Assembled machine, looking down into the developing, fixing, 
and stop-bath tanks. 



operator then removes the rack corresponding to the signalled num- 
ber. 

In Fig. 4 is shown a photograph of the assembled developing 
machine. At the extreme right on the shelf is the automatic timing 
mechanism. In the central part of the photograph is shown the 
superstructure of the developing machine, with the carriage upon 
which are hanging two racks filled with sensitometric strips. Im- 
mediately in front and at the center may be seen the top of the acid 



Jan., 1937] MACHINE FOR SENSITOMETRIC WORK 81 

stop-bath tank, and at the extreme lower left-hand corner may be 
seen a few film racks in the fixing tank. 

In Fig. 5 is shown another view of the assembled machine, taken 
looking down into the developing, fixing, and stop-bath tanks. At 
the rear may be seen a part of the double control mechanism which 
will be mentioned in the next paragraph. 

The temperature of the developing solution is held constant to with- 
in approximately =*=0.10F by means of a controlled water bath that 
surrounds the developing and stop-bath tanks. Warm and cold 
water lines (the former at about 90F and the latter at about 45F) 
furnish water to the bath for maintaining a constant temperature. 
The flow of water from these lines is controlled by two Sylphon 
water-mixing valves, actuated by means of the vapor pressure of 
highly volatile liquids contained in a copper bellows. The bellows 
is placed in the water bath to be controlled. Small variations in 
temperature of the bath cause large changes in the action of the valve 
as a result of the change in the vapor pressure of the liquid within 
the bellows. Two such valves are used, one on the warm water supply, 
the other on the cold. The settings of these valves are made to over- 
lap in such a way that both of them are always allowing water to 
enter the tank when the bath is at the normal temperature. The 
incoming water is circulated about the tank by means of baffle plates. 
The overflow passes through an outlet pipe to the washing tank and 
thence into the rinsing tank, and from there to the drain. The flow 
of water through the washing tank is sufficient to keep the hypo 
content low, but since the cross-section of the tank is rather large, 
the agitation produced by the incoming water from the water bath is 
not sufficient. Additional agitation in the washing tank is obtained 
by the injection of air at the bottom of the tank through a series of 
fine holes. 

Before starting to use this machine for our routine sensitometric 
testing, it was, of course, necessary to make an elaborate series of 
tests to establish its suitability for the work. It is essential that the 
same sensitometric characteristics shall be obtained for identical 
samples of photographic materials when processed in any one of the 
sixty positions in the machine, and also that the results shall be iden- 
tical, irrespective of whether the maximum exposure end of the strip 
be placed at the top or at the bottom of the tank. Moreover, we 
must be sure that identical samples of a photographic material proc- 
essed at widely different times shall give identical results. 



2 JONES, RUSSELL, AND BEACHAM [J. S. M. P. E. 

In order to make a thorough study of the effects of agitation upon 
the results, five different agitation conditions were studied, as follows: 

(a) No agitation. 

(6) Propeller only. 

(c) Paddles only. 

(d) Propeller and paddles (low speed). 

(e) Propeller and paddles (high speed). 

For this series of tests a quantity of motion picture positive film 
was used, carefully selected from a relatively old coating that had 
been seasoned for a long time in an atmosphere having a temperature 
of 70F and a relative humidity of 50 per cent. Great care was 

TABLE I 

Typical Data Obtained for Development with No Agitation 



Step 


6 


12 


24 


Position 
38 


44 


50 


56 


For 30 Strips 
Average Spread 


1 












. 








2 


2.66 


2.63 


2.67 


2.64 


2.66 


2.65 


2.66 


2.66 


0.04 


3 


2.48 


2.48 


2.51 


2.47 


2.49 


2.48 


2.48 


2.48 


0.04 


4 


2.34 


2.32 


2.34 


2.32 


2.34 


2.35 


2.30 


2.30 


0.05 


5 


2.16 


2.18 


2.20 


2.17 


2.17 


2.20 


2.18 


2.18 


0.04 


6 


1.97 


1.98 


2.01 


1.96 


1.98 


2.01 


1.97 


1.98 


0.06 


7 


1.79 


1.80 


1.82 


1.78 


1.80 


1.80 


1.79 


1.79 


0.04 


8 


1.59 


1.60 


1.62 


1.59 


1.62 


1.64 


1.61 


1.61 


0.05 


9 


1.41 


1.41 


1.41 


1.40 


1.41 


1.44 


1.42 


1.42 


0.04 


10 


1.21 


1.21 


1.21 


1.18 


1.21 


1.25 


1.21 


1.21 


0.07 


11 


1.00 


0.98 


1.00 


0.98 


1.00 


1.03 


1.00 


1.00 


0.05 


12 


0.79 


0.80 


0.82 


0.80 


0.82 


0.82 


0.81 


0.81 


0.03 


13 


0.63 


0.62 


0.63 


0.62 


0.62 


0.64 


0.63 


0.63 


0.02 


14 


0.45 


0.44 


0.45 


0.45 


0.47 


0.47 


0.45 


0.45 


0.03 


15 


0.32 


0.31 


0.32 


0.31 


0.32 


0.31 


0.31 


0.31 


0.02 


16 


0.20 


0.19 


0.21 


0.19 


0.21 


0.21 


0.21 


0.20 


0.02 


17 


0.12 


0.12 


0.12 


0.11 


0.12 


0.11 


0.12 


0.12 


0.01 


18 


0.06 


0.06 


0.06 


0.06 


0.06 


0.06 


0.06 


0.06 


0.00 



Maximum exposure end down. 

exercised to be sure that the material represented a maximum of 
uniformity throughout. Exposing the test strips was done on a 
Type lib sensitometer maintained under the usual invariant condi- 
tions. In loading the racks, each alternate strip was placed with the 
heavy exposure end downward, and the remainder with the heavy 
exposure end upward. In the first test, no agitation, the machine 
was filled with a carefully standardized mix of D-16 developer brought 



Jan., 1937] MACHINE FOR SENSITOMETRIC WORK 83 

to a temperature of 65 F 0.10, and allowed to stand for sometime 
so as to come to a perfectly quiescent state. The carriage was then 
lowered very carefully so as to create a minimum of disturbance 
within the body of the developer, and the entire sixty strips were 
allowed to develop for a period of five minutes. No agitation what- 
soever was given either to the developer or to the strips. 

After fixing, washing, and drying the strips, the densities were 
read visually on one of our standard densitometers. Some typical 
data obtained in this manner are shown in Table I. Here are given 
the readings made on seven selected strips which were developed 
with the maximum exposure end down. In the next to the last column 
will be found the average values for the entire thirty strips that were 
developed with the maximum exposure end down. The values in 
the last column are those of total spread within the entire group of 
thirty strips. It will be noted that these total spread values are 
exceedingly low, indicating that development has proceeded . with 
great uniformity for all thirty positions occupied by these strips in 
the tank. The data for the thirty strips developed with maximum 
exposure end up were treated in exactly the same way and the values 
for total spread were found to be of the same order as those shown in 
Table I. 

In Fig. 6 are plotted the density values obtained for the two groups 
of strips developed, one with maximum exposure upward and the 
other with maximum exposure downward. This shows very dearly 
the effect of the reaction products formed during development as 
they stream downward along the face of the exposed materials. This 
is ample demonstration that the condition of no agitation is one which 
can not be tolerated in sensitometric work. 

This procedure was then repeated exactly, with the exception that 
the solution was agitated by the operation of the propeller in the 
cylinder at the rear of the developing tank ; and exactly similar runs 
were made for the three other conditions of agitation indicated pre- 
viously, that is, with the paddles only, with the propeller and paddles, 
and with the propeller and paddles at accelerated speed. 

It does not seem necessary or desirable to attempt to present in 
detail all these data. The results can be satisfactorily analyzed for 
each of the conditions of agitation by a consideration of the data 
corresponding to those presented in the last two columns of Table I. 

In Table II are shown the average density values for the thirty 
strips developed with the maximum exposure end downward under 



84 



JONES, RUSSELL, AND BEACHAM [J. S. M. p. E. 



each of the agitation conditions. It is quite evident from a considera- 
tion of these values that there has been a progressive increase in the 
rate at which development proceeds. This rate of increase is very 
marked for the first three steps of increasing agitation, while running 
the paddles at accelerated speeds produces only slight increases in 
the density values obtained, and these increases are confined almost 
entirely to the higher densities. It will be noted that in the low- 

TABLE II 

Average Density Values for the Thirty Strips Developed under Each of the Agitation 

Conditions 



Step 


No 
Agitation 


Propeller 


Paddles 


Propeller & 
Paddles 
(Low Speed) 


Propeller & 
Paddles 
(High Speed) 


2 


2.66 




. . 






3 


2.48 




. . 




. . 


4 


2.30 






. . 




5 


2.18 


2.89 


3.01 


. . 




6 


1.98 


2.69 


2.84 


2.94 


3.02 


7 


1.79 


2.49 


2.63 


2.73 


2.79 


8 


1.61 


2.27 


2.39 


2.49 


2.55 


9 


1.42 


2.04 


2.13 


2.24 


2.28 


10 


1.21 


1.76 


1.84 


1.95 


1.98 


11 


1.00 


1.48 


1.55 


1.65 


1.67 


12 


0.81 


1.19 


1.25 


1.33 


1.33 


13 


0.63 


0.90 


0.95 


1.01 


1.01 


14 


0.45 


0.63 


0.68 


0.72 


0.72 


15 


0.31 


0.43 


0.45 


0.49 


0.50 


16 


0.21 


0.26 


0.28 


0.30 


0.30 


17 


0.12 


0.15 


0.16 


0.16 


0.16 


18 


0.06 


0.08 


0.09 


0.10 


0.10 


Speed 


12.9 


14.6 


15.1 


16.6 


16.6 


Gamma 


1.48 


1.86 


1.94 


2.01 


2.05 



Values in above taken from the average of all strips developed with maximum 
exposure end down. 



density range, the agreement between low-speed and high-speed 
paddle operation is excellent. At the bottom of the table are found 
the values of speed and gamma computed from the curves plotted 
from average densities, and, as would be expected, both these values 
increase as agitation is increased, representing, however, practical 
equilibrium for the last two conditions. 



Jan., 1937] 



MACHINE FOR SENSITOMETRIC WORK 



85 



In Table III are tabulated the total spread values for the five 
conditions of agitation. These values show some very interesting 
facts. As mentioned previously, the total spread values obtained 
with no agitation are remarkably low, indicating excellent uniformity 
of development from position to position in the tank. With the pro- 
peller operating, these total spread values rise to an astonishing 
extent, indicating that with this amount of agitation, the lack of 



TABLE III 

Total Density Spread Values Obtained for the Five Conditions of Agitation 



Step 


No 
Agitation 


Propeller 


Paddles 


Propeller & 
Paddles 
(Low Speed) 


Propeller & 
Paddles 
(High Speed) 


2 


0.04 


.. 


.. 






3 


0.04 


. 








4 


0.05 










5 


0.04 


0.41 


0.10 






6 


0.06 


0.42 


0.10 


0.08 


0.04 


7 


0.04 


0.34 


0.10 


0.09 


0.05 


8 


0.05 


0.39 


0.07 


0.07 


0.04 


9 


0.04 


0.36 


0.06 


0.05 


0.03 


10 


0.07 


0.28 


0.05 


0.05 


0.04 


11 


0.05 


0.29 


0.04 


0.04 


0.03 


12 


0.03 


0.27 


0.07 


0.02 


0.02 


13 


0.02 


0.15 


0.03 


0.03 


0.04 


14 


0.03 


0.15 


0.04 


0.04 


0.02 


15 


0.02 


0.12 


0.05 


0.03 


0.02 


16 


0.02 


0.05 


0.04 


0.02 


0.02 


17 


0.01 


0.03 


0.02 


0.02 


0.02 


18 


0.00 


0.02 


0.01 


0.02 


0.02 



Values in above table taken from average of all strips developed with maximum 
exposure end down. 



uniformity from point to point in the tank has increased enormously. 
This is quite readily understood, however, when we consider that 
the agitation is in a single direction and that the currents may move 
in fairly fixed paths which may be far from straight lines, thus pro- 
ducing a more or less definite pattern of developer replacement over 
the surfaces of the exposed strips. It is obvious that with this 
amount and this kind of agitation the uniformity of development 
from point to point in the tank is extremely poor. With the paddles 



86 



JONES, RUSSELL, AND BEACHAM [J. S. M. p. E. 



AJ.ISN3Q 

**S225S8 a? ' i? 2^ p *2S'S 


O I.O Z.O O I.O "20 

LOG E. LOG C. 

FIG. 6. (Left) Characteristic curves plotted from the averages of two groups of strips developed with no 
agitation. 
FIG. 7. (Right) Characteristic curves plotted from the averages of two groups of strips developed with agita- 




























r 
































AXIMUM EXPOSURE: UP 


i. 

O 

n 






























y. 












V 






















- 












^ 


















o; 
















^ 


v 

' 














\ 




















s 


9^ 


































S 








* 


























^ 


s 






i ( 

- o A 




























n "S 


K 




































N 


v 


































X 


\ 


































> 


Si 


































\ 




























































































= 1 

\ 

< 

\it 

I , 

X. 










\ 


\ 
































N 


k 


































V 
















AGITATION 


















^x 






























> 


^x, 


X. 
































X 










Z 
























N 


. \ 

s. 


N 






























^ 


T 


































'\ 




b 
































x 


^ 


\ 


































X 


V 


































































1 




































<VOCO>NpcO<, N ocB M 
<rtrt^NNN(MN----_:bSdo 
AJ.ISN3Q 



Jan., 1937] MACHINE FOR SENSITOMETRIC WORK 87 

operating alone the condition is greatly improved, but still the total 
spread and therefore the nonuniformity is much greater than that ob- 
tained with no agitation. 

With the propeller running and the paddles operating at the lower 
speed, further improvement is obtained, but the uniformity is still 
not as good as that obtained with no agitation. By increasing the 
speed of the paddles, further improvement occurs, and now the total 
spread is of the same order as, or perhaps slightly better than, that ob- 
tained with no agitation. These total spread values are so low that 
we feel quite confident in saying that a satisfactory state of uniformity 
in the developing tank has been reached. In other words, the devel- 
oper is moved so rapidly and so thoroughly that every point of the 
surface of the exposed strips is subjected to the action of a developer 
that is uniform in power, and the reaction products are being swept 
away from these surfaces at such high velocities and so completely 
that there is no nonuniformity of development due to localized high 
concentrations of reaction products. 

We wish now to turn to a further consideration of the results ob- 
tained from the groups of strips developed in part with maximum 
exposure upward and in part with maximum exposure downward. 
These results for the condition of no agitation have already been 
shown in Fig. 6. In Fig. 7 are shown the two characteristic curves 
plotted from averages of the two groups of thirty strips developed 
by agitation with propeller only. It will be seen that there is a slight 
but definite difference. In Fig. 8 it will be seen that the two curves 
representing the two groups of thirty strips have now become coinci- 
dent, and the same is true for the other two conditions of agitation, 
as illustrated in Figs. 9 and 10. 

An inspection of the data in Table II has already revealed that as 
the agitation is increased in violence the rate of development has 
increased appreciably. In order to show this more graphically, the 
curves in Fig. 11 are presented. It will be seen here that for the 
two conditions of maximum agitation there is still a slight difference, 
especially in the region of high densities (D greater than 1.8). The 
change, however, is so slight that we felt justified in assuming that 
further increase in agitation did not produce appreciable increase 
in the magnitude of density resulting from a given exposure ; and we 
have concluded that, for the work to be done with this equipment, 
using the agitation produced by the paddles at the accelerated velocity 
supplemented by the propeller, a sufficiently high rate of development 



88 



JONES, RUSSELL, AND BEACHAM [J. S. M. p. E. 



<ti tS 



I L: - 6 





























2 








rt ^ 

p -5 1 
























ffi 




$1 








> - 
























U 

_J 
n 




U 


i 






1 1 




























u 


2. 








a ^ 
*a> % 






\ 


















o 

r 


u 


a 

X 








1 f 








X 
















4 


j 


u 








^ t 

C/2 

Q u cx *o 










*" 


x, 












_ I 


2 


D 

y 








oS a' 2 














^ 


^v 








U 




^ r 








J ^ ^^ 

O o en 


















X) 






a 




1 


| 






a &* 
_ o 
























^ 


^ 










s i- 

bo -^ cd 




























^ 








o i ^^ 






























*^ 


s, 




-*-* t/5 J 

*S w>2 


































1 


>^ rt QJ 
3^ fc 


































5 


bO > 

2 s rt cS 




































>l^^ 

rf ^ *5 c 
































































2 








^aS a 

P -M ^ O T) 




























5;c 








" i!^ 1 




























u 








*lli 




> 


1 




















J j 


3 
</> 








1 |I 

4_> o< o 








\ 


















<fl 


r 








o 8 

"a a> a 










k^ 


















u 








obi | ^^ 












V, 


^ 












< 


2 
D 








"o ol 
















^ 


\ 










>< 
< 








J .0 ;^g 




















V 
















" a! So 
S o rt 
























X 


^ 










^ rt 
ed o3 




























^ 








^ 
_ _g O 






























* 


N 




S J 




































^ -^ 
V* r5 




































S * 




































00 ^ 



Jan., 1937] 



MACHINE FOR SENSITOMETRIC WORK 



has been attained so that the limit of density obtainable from a given 
exposure has been closely approximated. 

In practical work the characteristics of a given emulsion are 
generally determined by developing a relatively few strips. Thus, 
it is of utmost importance to know, with a high degree of certainty, 
whether there is any difference in the extent of development of strips 
developed in one position in the machine as compared with those in 
any other position. To establish this definitely, five complete runs 
were made under the condition of maximum developer agitation. 
The density values from each of the five strips for a single position 
in the machine were then averaged, and these averages for the sixty 
positions compared with each other. Again, we shall not present 

TABLE IV 

Average Density Values and Total Density Spread of Strips Developed 
in Different Positions in Developing Machine 



Step 


Average 


Spread 


6 


3.03 


0.03 


7 


2.79 


0.04 


8 


2.55 


0.03 


9 


2.29 


0.02 


10 


1.99 


0.03 


11 


1.67 


0.02 


12 


1.34 


0.02 


13 


1.03 


0.03 


14 


0.72 


0.02 


15 


0.49 


0.02 


16 


0.30 


0.02 


17 


0.16 


0.01 



the data in detail, but shall illustrate the results by giving the average 
density values together with the total spread. These are shown in 
Table IV. It will be seen that the values of total spread are ex- 
tremely small when it is considered that even these can not necessarily 
be ascribed entirely to nonuniformity of processing, since that may 
include certain errors due to reading and other factors. 

Since in the past, hand-rocked and mechanically rocked trays have 
been used for the work that is now being done with this developing 
machine, it is of interest to compare the rate of development attained 
by the two methods. This information is of interest in drawing con- 
clusions as to the development times that must be used with the 
machine method to give extents of development comparable with the 



90 



JONES, RUSSELL, AND BEACHAM 



[J. S. M. P. E. 



AJLISN3Q 


p 
N 

olii 

_i 



o 

N 

2ui 

(5 

3 
































b 3- PADDLES ONLY (LOW SPEED) 
4-PROPELLER AND PADDLES 1 c 
(LOW SPEED) 


1 
> 








N 


















or 



y I-NO AGITATION 
r 2- PROPELLER ONLY 




k\N 


^ 




\ 
















> 


L\ 




\ 


\ 
















Ul 




\ 


vs 






s 


\ 












tf 


r U 








^N> 


<^ 


x 




X 










U 


if 










^ 


^ 


>> 




N 


X 






^ 














^ 


CD^NI 


^ 




x, 






\ ~ 

i 


















^ 


N 


^> 


N 
























^ 


^ 


^ 






























^ 


X 


































^ 


^ 


































1 


^ 


































\ 




























































































r 






























PROPELLER AND PADDLES 
(HIGH SPEED) 

_0- MAXIMUM EXPOSURE Uf 


I 




























































= 








^ 


^ 
































V 
































> 


N^ 














= 


















^ 


3^ 
































^ 


x^ 






4 


























^^ 


X 




































^ 


X 




































^ 


S. 


































> 









































































Jan., 1937) 



MACHINE FOR SENSITOMETRIC WORK 



91 



older technic. For this purpose one of our machine-rocked tray equip- 
ments was used. In Fig. 12 the results obtained are shown graphi- 
cally. Curves for the density of five selected steps are plotted, density 
being shown as a function of frequency of rocking. It will be seen 
that the density obtained by this method rises rapidly in the region 



4.0 

38 
36 
34 

30 
2 6 
26 



\Z ROCKS 



MIN 



.ATM. STEP 



TTM. STEP 



h 

, 

g.. 



I 
08 
0.6 



IOTH- STEP 



I ATM. 



STEP 



iaTH. STEP 



O 2 4 6 8 IO \Z 14 16 18 ZQ ZZ 24 
ROCKS/ M)N 



FIG. 12. Results obtained with mechanically rocked 
tray equipment. Density plotted as a function of fre- 
quency of rocking. 



of low frequency of rocking to a limiting value which becomes con- 
stant in the region of higher frequency. The dotted horizontal lines 
represent the densities obtained with the developing machine for 
the same time of development, using developers of identical com- 
position and temperature. It will be seen from this that the develop- 
ing machine is giving an extent of agitation corresponding to the 



92 



JONES, RUSSELL, AND BEACHAM [J. S. M. p. E. 



maximum obtainable in the mechanically rocked, and, by inference, 
in the hand-rocked technics. 

One other method of developing for sensitometric purposes has 
been recommended and used to some extent. This is the so-called 
"brush method" of development originated by Clark 1 and later used 
and elaborated to some extent by other workers in this field. In this 
method, agitation is produced by using a soft, camel's hair brush 
which is swept by hand continuously back and forth across the emul- 
sion surface. It has been stated that this results in a maximum 

TABLE V 

Results Obtained by Brush Development Compared with Those Obtained by Machine 

Development 



Step 


C.M.K. 


H.R.B. 


Brush 
W.A.S. 


Average 


Spread 


Machine 
Average Spread 


2 


3.36 


3.34 


3.40 


3.37 


0.06 


3.35 


0.04 


3 


3.30 


3.28 


3.35 


3.31 


0.07 


3.30 


0.05 


4 


3.24 


3.21 


3.31 


3.25 


0.10 


3.25 


0.03 


5 


3.11 


3.10 


3.18 


3.13 


0.08 


3.10 


0.04 


6 


2.97 


2.94 


3.05 


2.99 


0.11 


2.93 


0.03 


7 


2.79 


2.75 


2.90 


2.81 


0.15 


2.73 


0.04 


8 


2.56 


2.54 


2.60 


2.57 


0.06 


2.49 


0.03 


9 


2.31 


2.28 


2.41 


2.33 


0.13 


2.24 


0.02 


10 


2.04 


2.02 


2.12 


2.06 


0.10 


1.97 


0.03 


11 


1.74 


1.71 


1.81 


1.75 


0.10 


1.67 


0.02 


12 


1.43 


1.40 


1.49 


1.44 


0.09 


1.36 


0.02 


13 


1.15 


1.10 


1.19 


1.15 


0.09 


1.07 


0.03 


14 


0.86 


0.78 


0.87 


0.83 


0.09 


0.77 


0.02 


15 


0.60 


0.55 


0.61 


0.59 


0.06 


0.55 


0.02 


16 


0.39 


0.36 


0.38 


0.38 


0.03 


0.35 


0.02 


17 


0.23 


0.22 


0.21 


0.22 


0.02 


0.20 


0.01 


18 


0.14 


0.13 


0.14 


0.14 


0.01 


0.13 


0.00 


19 


0.08 


0.07 


0.07 


0.07 


0.01 


0.07 


0.00 



possible rate of development and in maximum uniformity of develop- 
ment over the surface of the materials being treated. Our experience 
with this method has indicated that it does give excellent uniformity 
over a given portion of material, but that different operators may 
differ appreciably from each other in the amount of development 
obtained in a given time. It seemed desirable to compare the develop- 
ment produced by this machine with that produced by the brush 
technic; and for this purpose a test was run in which three different 
operators applied the brush method to a group of strips, and a similar 



Jan., 1937] 



MACHINE FOR SENSITOMETRIC WORK 



group of strips was run in the machine. The results are shown in 
Table V. In the first three columns are the values obtained by the 
individual operators using the brush technic. In the next columns are 
the averages of these developments, together with the spreads in 
density obtained among the various developments. In the last 
columns are the values obtained by the machine processing together 
with the spreads normally found for this method of development. 
It will be seen that the spread in results obtained among the opera- 

TABLE VI 

Values of Speed and Gamma Obtained from Developments Made on Different Days 
over a Period of Approximately One Month 



Tested 


Speed 


Deviation 


Gamma 


April 20 


20.0 


-0.27 


2.05 


24 


20.4 


+0.13 


2.07 


28 


20.6 


+0.33 


2.06 


30 


20.4 


+0.13 


2.05 


May 2 


20.0 


-0.27 


2.04 


3 


20.2 


-0.07 


2.03 


3 


20.6 


+0.33 


2.09 


4 


20.4 


+0.13 


2.08 


5 


20.0 


-0.27 


2.05 


7 


20.2 


-0.07 


2.03 


8 


20.0 


-0.27 


2.02 


9 


20.0 


-0.27 


2.02 


12 


20.6 


+0.33 


2.08 


13 


20.4 


+0.13 


2.02 


Average 


20.27 * 


1.05% 0.21 


2.05 =*= 1.0% 


Spread 


0.60 




0.07 


Spread 


3% 




3.4% 



tors using the brush method is much greater than that found by 
machine processing. The density values in the "average" column 
for the three brush developments are slightly higher than those ob- 
tained by the machine, indicating that the brush technic probably 
does give somewhat more perfect removal of the reaction products 
from the surface of the materials being developed than does the ma- 
chine. The difference, however, is not great and in view of the 
spread between different operators it appears that the machine tech- 
nic is definitely superior when the volume of work is such as to re- 
quire several different operators. 

Finally, as an indication of the time uniformity of the work that 



94 JONES, RUSSELL AND BEACHAM [J. S. M. p. E. 

can be done with this developing machine, the values shown in 
Table VI are presented. These show the values of speed and gamma 
obtained from runs made on different days over a period of approxi- 
mately one month, and are derived from the average of strips all 
cut from a single roll of motion picture positive film carefully aged 
and seasoned to represent a maximum of uniformity throughout its 
length. It will be seen that the mean deviation from the average 
speed is only =*= 1 per cent, while the mean deviation from the average 
gamma is also =*= 1 per cent. The maximum spread in the case of 
speed is 3 per cent, while that for gamma is 3.4 per cent. It must 
be remembered that these deviations may be attributable to several 
factors other than the lack of reproducibility and uniformity of 
agitation and temperature control within the developing machine. 
They certainly include possible variations in the concentration and 
constitution of the developer and possible errors in the determination 
of density. 

We feel confident in stating that the performance of the develop- 
ing machine is all that can be desired, and that the errors that 
formerly might have been attributable to lack of uniform and re- 
producible agitation in the processing of sensitometric strips have 
now been reduced to a point far below the errors attributable to 
other causes. 

REFERENCE 

1 CLARK, W.: "Standard Development," Phot. J., 49 (Feb., 1925), No. 2, p. 

76. 

BIBLIOGRAPHY 

SHEPPARD, S. E., AND MEES, C. E. K. : "Investigations on the Theory of the 
Photographic Process," Longmans Green & Co. (London), 1907. 

TADAROKU, O.: "On a Developing Apparatus with a Rotary Plate Holder," 
Bull. Kiryu Tech. College, 1921. 

SHEPPARD, S. E., AND ELLIOTT, F. A.: "The Influence of Stirring on the Rate 
and Course of Development," /. Franklin Institute, 195 (Feb., 1923), No. 2, p. 
211. 

SHEPPARD, S. E., AND ELLIOTT, F. A. : "The Influence of Stirring on the Rate 
and Course of Development Part II," Ibid., 198 (Sept., 1924), No. 3, p. 333. 

HEPWORTH, C. M.: "A Developing Machine," Brit. J. Phot., 74 (May, 1927), 
No. 5, p. 260. 

BLOCH, O., AND HORTON, H. : "A Mechanical Developing Appliance for Sensito- 
metric Work," Phot. J., 52 (Aug., 1928), p. 352. 

SHEPPARD, S. E., LAMBERT, R. H., AND AKINS, G. M.: "Semi-Automatic De- 
velopment of Plates in Sensitometry," IX Internal . Congress Sci. and Applied 
Phot.; Revue d'Optique Theorique et Instrumental. 






Jan., 1937J MACHINE FOR SENSITOMETRIC WORK 95 

DISCUSSION 

MR. TOWNSLEY: You spoke of reaching a flat part of the speed curve. In 
what terms is the speed expressed? 

MR. JONES: I meant the flat part of the agitation curve. The curves repre- 
sent growth of density with the cycles per minute of rocking the tray. The 
speed values were obtained for five different conditions of agitation. For the 
first one, with no agitation, the speed value was relatively low. Then it increased 
and became 16.6, in the next to the last one. The last one was 16.6, which means 
that increasing the agitation of the next to the last did not produce any increase 
in speed. We therefore assumed that we had reached the flat part of the speed 
curve, speed being plotted as a function of the agitation. 

MR. TOWNSLEY : Has that been carried far enough to show whether the curve 
dropped off again with very, very high agitation? 

MR. JONES: I believe that Dr. Sheppard has made some experiments at ex- 
tremely high agitations. I am not sure what the results were. 

MR. SHEPPARD : I do not know how the speed would compare with the agita- 
tion in your case, and in rotations I would not say what it was. With the de- 
velopers that we used, the density over the whole of the curve, as a function of 
agitation, was a function of the number of revolutions of a propeller. This den- 
sity went through a rather flat maximum. I do not know to what to attribute 
that- I do not see exactly why it should not stay perfectly flat, but it seemed to be 
a perfectly evident phenomenon. 

MR. JONES: Might we assume that you had the developer moving so fast that 
it did not have time to stop and do its work? 

MR. SHEPPARD: It might be possible. There is another possibility; we found 
in certain cases that the effect of bromide was altered by the velocity of agitation, 
MR. OSWALD: In obtaining these data, have you worked out the smallest 
permissible amount of developer in relation to the emulsion surfaces being de- 
veloped? 

MR. JONES: We have not carried that to an ultimate conclusion. This tank 
holds about 7 l / 2 gallons of developer and accommodates 60 strips. After having 
used it for one development, if we reload and develop another 60 strips in that 
same developer there is perceptible change. For work of highest precision we 
can not use a developer more than once. In certain work in which the require- 
ments are not quite as exacting we do use it twice. But for routine testing in 
which we want the maximum of precision, we always discard the developer 
after having developed 60 strips. 

MR. FARNHAM: For a person doing sensitometric work without such equip- 
ment, your data indicates that 12 rockings per minute, or even a little faster 
might be satisfactory. Also, is the rocking of the tray continuous motion? 

MR. JONES: I have no sufficiently precise specification of the gear mecha- 
nism that was used to rock the tray to say exactly what the motion would be. 
I assume it is something approximating a simple harmonic. It is probably pro- 
duced by a rotating wheel and a lever. 

You ask whether one can rely upon that particular number of rocks per minute 
to be satisfactory. No, because the agitation is produced at the surface of 
the film and might be very different with a different mechanical structure. If 
you built yourself a mechanically rocked tray, you would have to run just such a 



96 JONES, RUSSELL, AND BEACHAM [J. S. M. p. E. 

series as we did, increasing the cycles per minute, plot the growth of density with 
frequency, and then find out where the curve flattens. 

MR. MILI: Can you define the speed as used in these various tables accu- 
rately? 

MR. RUSSELL: The values in the table were determined from the exposure 
required to produce a certain density. The density was 0.30 in this case. 

MR. JONES: In the early part of the paper I mentioned brush development. 
That technic has been recommended by a good many workers. It does produce 
very uniform development over the area of a single sample of film, and in the 
hands of a single operator it affords very satisfactory reproducibility from time 
to time. But, as I stated, in the hands of different workers we have found con- 
siderable variation. We used three operators, who were given instructions and 
told to brush at the rate of so many cycles per minute. They tried to follow the 
instructions. The results of the different operators showed about two to three 
times as much spread as the machine. 

The explanation is, of course, that perhaps even if they do carry out the proper 
technic, one man will press a little harder on the brush than another, and it just 
seems impossible with any human manipulation to make eight or ten different ob- 
servers produce the same results ; by which I mean results of the same uniformity 
as we can get with mechanically driven agitating systems. 

MR. FAMULENER: You showed very little deviation between strips. Would 
you say that by using no agitation it would be possible to develop strips that, while 
they would not reach their full development, would, if the same technic were fol- 
lowed as regards heavy end down or heavy end up, be entirely comparable 
throughout a series of tests? 

MR. JONES: The evidence indicates that. It is extremely bad practice, how- 
ever, for the reason that you have impressed upon the photographic material a 
series of exposures, and are going to plot the results in density as a function of the 
exposure. Of course, you will want to know later how the material is rendering 
different amounts of exposure. 

If you mount the strip, we shall say, with the heavy exposure end up, reaction 
products will be produced that will stream down over the other areas of the sen- 
sitometric strip, giving a distorted value. Thus, processing with no agitation 
will show you the characteristics of the material when developed only under the 
particular conditions. I do not see how you can apply it to other work where that 
same condition of complete quiescence does not exist. I think it is very bad 
practice to do it. 

MR. SCHMIDT: What is the minimum development time necessary to achieve 
uniform results? In other words, can a short working development be applied 
safely, or do you prefer to develop for at least five or ten minutes? 

MR. JONES: The minimum development time will depend upon the developer, 
upon the photographic material, and upon the degree of agitation. If we should 
substitute for the developer that we were using, which, by the way, was D-16, 
a very active developer such, let us say, as a caustic type, we should get comple- 
tion of development in a very much shorter time with the same agitation. 

MR. POTTER: What are some of the precautions necessary to determine the 
identity of composition of the developer being used from time to time during a 
series of experiments? 



Jan., 1937] MACHINE FOR SENSITOMETRIC WORK 97 

MR. JONES: We purchase the ingredients in relatively large quantities and of 
the maximum obtainable purity. I think that Mr. Crabtree can tell you what 
tests he applies for raw materials. Then, of course, the raw materials are mixed 
by one of Mr. Crabtree's assistants, the greatest of care being given to weighing, 
mixing, and all that sort of thing. Then we blend. We have a system in which 
we use several different mixes: assuming that we have ten available mixes, we 
take one gallon out of each mix. That is the general scheme, although the 
numbers quoted may not be exact. 

After having the final developer, we give it a sensitometric test, using a sample 
film that is carefully conditioned, aged, and laid aside for the purpose. If we 
get a batch that shows variation in contrast or in speed greater than a certain 
tolerance that has been established by experience, it is rejected. 

MR. CRABTREE: We take as large quantities of each ingredient as possible, 
and treat them as standards for the sensitometric work. When the stocks are 
almost exhausted we make a careful photographic check of the new stock against 
the old stock. 

Answering Mr. Schmidt, the disadvantage of rapid development is, of course, 
the possibility of errors involved in transferring the film from the developer to 
the stop bath. As the time of development is extended the error diminishes pro- 
portionately. 

MR. MATTHEWS: In precision work have you any minimum number of strips 
that you average for best results? 

MR. JONES: It is impossible to generalize. The number of strips de- 
sirable for averaging will depend, to some extent at least, upon the type of photo- 
graphic material. It is desirable, of course, to have a development technic so 
uniform, so reproducible, that the number of strips to be averaged can be reduced 
to a minimum. 

A study of the data in the paper will show that the total spread in the values 
derived from a large number of strips developed by the machine technic is much 
lower than the total spread within an equal number of strips developed by the 
hand-rocked tray technic. That, of course, means that to obtain a result subject 
to some specified probable error, the number of strips that must be averaged in 
the case of the machine technic will be definitely less than in the case of the hand- 
rocked tray technic. I do not have data at hand showing exactly the relation- 
ship between the number of strips to be averaged in the two cases to give the same 
probable error. 

MR. MARCUS: Do you intend to use a mechanical device for dipping the film 
into the developer and lifting it out, so as to eliminate the human factors? 

MR. JONES: No, we think that is unnecessary. The operation is so simple 
and the time consumed in doing it is so small compared to the time of develop- 
ment that we feel sure the variation due to this operation is quite negligible. 

It is possible that it might be necessary if one were using development times 
and conditions such that at the time of removal of the strip development was 
{proceeding at a high rate; but if we are up pretty well to the equilibrium point 
then, of course, errors due to that variation in time would be very small. 
i MR. SCHMIDT: Have the sensitometer strips used to prepare the slides that 
have been shown, all been dried under definite conditions? 

MR. JONES : Yes, the drying conditions throughout were constant. 



98 JONES, RUSSELL, AND BEACHAM 

MR. TOWNSLEY : The curves show that less time is required to obtain a given 
gamma with greater agitation. Have you carried the work far enough to reach 
the peak? 

MR. JONES: No. I tried to make clear the fact that with lower degrees of 
agitation the gamma was growing quite rapidly. Then, for the maximum 
agitation and the step just next to the maximum of those shown there was a 
change of only 0.04 in gamma. The time rate of change in gamma became very 
low. 

MR. TOWNSLEY: You find the whole time-gamma curve shifting upward? 

MR. JONES: Yes. 



NOTE ON THE USE OF AN AUTOMATIC RECORDING 
DENSITOMETER* 



C. M. TUTTLE AND M. E. RUSSELL** 



Summary. A recording physical densitometer designed to read strips from the 
type lib sensitometer has been previously described. 1 This instrument has been in 
service in the sensitometric department of the Kodak Research Laboratories for about 
one year, during which time it has been operated steadily. Approximately 100,000 
sensitometric strips have been read thus far. The instrument is capable of an output 
of about 700 strips per day. 

Experience has shown that more tepeatable results are attained with this instru- 
ment than by routine, visual methods. Comparative data accumulated in an ex- 
periment lasting several weeks are presented, together with a time study of the two 
methods of densitometry . Certain features to be changed in the design of a new in- 
strument will be discussed. The new instrument will be improved in both ruggedness 
and speed. 

The advisability of using devices of this nature in a release print laboratory will de- 
pend upon a number of factors, such as initial cost, quantity and quality of output, 
and ease of maintenance. 

A recording physical densitometer designed to handle routine 
work for the sensitometry department of these Laboratories has been 
in continuous service for about a year. During this time a consider- 
able amount of data concerning its speed, accuracy, and durability 
has been obtained. It is the purpose of this paper to compare these 
factors with corresponding data for routine visual methods. 

DESCRIPTION OF THE INSTRUMENT 

Since the principles of design and operation of this instrument 
were discussed at some length in a previous paper 1 an abbreviated 
description taken from that paper will suffice for the present purpose. 

The instrument operates on the two-cell null method, with con- 
trolled intensity of the measuring beam. The schematic drawing, 
Fig. 1, shows the essential features of the apparatus. The lower 

* Received October 9, 1936; Communication No. 603 from the Kodak Re- 
search Laboratories. Presented at the Fall, 1936, Meeting, at Rochester, N. Y. 
** Eastman Kodak Co., Rochester, N. Y. 

99 



100 



C. M. TUTTLE AND M. E. RUSSELL [J. S. M. P. E. 



light-beam illuminates the comparison cell, and the intensity of this 
illumination may be manually regulated within a small range to set 
the instrument zero. Light in the second beam traverses an aper- 
ture variable as to area over a maximum to minimum ratio of 1000 
to 1 .0. In the drawing, the two beams are shown as if they originated 
from different sides of the lamp. The actual optical system makes use 
of a beam-splitter. After passing through the variable aperture, the 
light is incident upon one of the areas of the sensitometric strip. 
Nearly in contact with the photographic material is placed the sur- 







FIG. 1. Schematic diagram of automatic densitometer. 

face of the measuring photocell. This part of the optical system 
follows the suggestion of a previous paper. 2 The two photocells, 
the one in the comparison beam and the other in the measuring 
beam, are connected in opposed potential to a galvanometer. 

Light from a second source collimated by a second optical system 
is reflected by the galvanometer mirror upon a large condenser lens. 
The lens, except for an open slit at the zero position of the galva- 
nometer, is covered by an opaque mask. At zero galvanometer de- 
flection, the light is concentrated upon a third photocell. The out- 
put of this cell is used to close a relay which initiates the operation 
of recording and shifting the coordinate paper and the sensitometric 
strip. 

In operation, the first step of the sensitometric strip is registered 



Jan., 1937) AN AUTOMATIC RECORDING DENSITOMETER 101 

over the window of the measuring photocell, thereby decreasing the 
light intensity upon the cell and throwing the photocell -galvanometer 
circuit out of balance. A hand-operated switch closes the circuit of 
a magnetic clutch, which drives the diaphragm-changing mechanism 
in the direction for increasing the light in the measuring beam. 

When the light intensity upon the measuring cell has become 
equal to that upon the comparison cell, the relay is tripped. The 
relay circuit energizes a solenoid pulling a plunger, which releases 
an over-running clutch and allows it to make one revolution. The 
over-running clutch drives a shaft upon which are cams operating 
two switches. The first small fraction of the turn of the shaft through 
the medium of these switches de-energizes the diaphragm-driving 
clutch and discharges a spark through the graph paper from a metal 
pointer. The pointer is attached to the end of the traveling dia- 
phragm; thus, its position at all times is a positive indication of the 
disclosed area of the variable aperture. At the time of the spark dis- 
charge, the sum of the effective density of the diaphragm in its optical 
system and the density of the material being measured is constant. 

Upon the shaft that is rotated by the over-running clutch is cut 
a screw, the lead of which is equal to the 0.15-increment upon the 
abscissa scale of the graph paper. As the shaft completes its single 
revolution, a stainless steel carriage table, upon which the graph 
paper is first manually registered and then held by suction, is moved 
forward the proper amount. Attached to the table is an arm which 
pulls the sensitometric strip the distance of one step. 

At the termination of the operations just described, the instru- 
ment has set itself for recording the second point. The diaphragm 
has been pulled backward enough to unbalance the photoelectric 
circuit even if the second step should be no greater in density than 
the first. The magnetic clutch that drives the diaphragm is energized 
as soon as the shaft completes its revolution, and the upward travel of 
the diaphragm continues until a match of the two cell intensities 
is again achieved. Upon the completion of the point-by-point record 
of all twenty-one steps, the driving circuit is automatically broken 
and the paper may be removed. 

SPEED OF OPERATION 

The densitometer has been designed to carry a maximal load of 
about seven hundred lib sensitometric strips per day with a single 
operator. The bulk of the work that it handles consists in plotting 



102 C. M. TUTTLE AND M. E. RUSSELL [J. S. M. P. E. 

families of curves of varying times of development for positive film. 
Eight sets of points for eight strips are customarily plotted upon a 
single sheet. For this type of work, the average time per strip is 
fifty seconds. Since, in routine visual densitometry of the same kind 
of work, the average reading rate is one strip in two minutes, with 
one person reading while a second person records, the machine method 
may be said to be nearly five times as fast as the visual method. The 
machine has the additional advantage that by being operated for 
two shifts instead of one, it can take care of peak loads without 
having to employ additional skilled photometricians. 

ACCURACY AND REPEATABILITY 

The results of a series of tests that show the accuracy and repeat- 
ability of measurements made by the physical densitometer in com- 
parison with measurements made by visual instruments will now be 
discussed. Typical sensitometric strips were chosen for a number of 
materials, including both negative and positive motion picture films 
with clear, gray, and colored bases. Each strip was read on ten 
occasions, covering a period of about 6 weeks. In order to obtain 
visual readings typical of those normally made in routine sensitome- 
try, the strips were mixed in with daily test-strips, and readings made 
by the regular operators in the normal manner. The visual measure- 
ments were made on an instrument using a Martens polarization 
photometer of which the optical characteristics are such that the 
results agree with those obtained by a Jones high-intensity den- 
sitometer. 3 

In Table I are typical results obtained for a gray-base motion pic- 
ture negative material. Each of the ten readings is tabulated to- 
gether with the averages for each step. Comparing the average 
values obtained by the two methods, it will be noticed that the results 
for the physical instrument are in practically perfect agreement with 
those obtained visually. For no step is the difference in density more 
than 0.01. 

When the individual values themselves are examined, it is at once 
apparent that the physical instrument gives noticeably more con- 
sistent results than those attained by the visual method. For ex- 
ample, on the twelfth step all the values obtained by the physical 
instrument are either 0.63 or 0.64. When read visually, the values 
range from 0.61 to 0.65. In some cases the range of visual values is 
somewhat less than that and in a few cases somewhat larger. 



Jan., 1937] AN AUTOMATIC RECORDING DENSITOMETER 



103 






Average 


^rHOO^OiCOOl^-OO 

coeqooii>coioco<N 
r_irH'rH'dddddd 


rHrHrHOOOddd 


ooo 


sasssssfea 


CO C^ *O O^ t^* CO *O CO CO 


C/3 


rHrHrHOOOOOO 


rHrHrHOOOOOO 


** 


CO C^ ^^ Oi t^* CO *O CO CQ 


Tf(NOiOO5iO<NOOOO 
CO C^ ^^ Oi t^* CD *O CO C^ 


3 *" 


rHrHrHOOOOOO 


rHrHrHOOOOOO 


I 1 ' 


lOrHt>-COOCOOt>t>- 

CO<NOOiOOcOiOCO<N 

rHrHrHOOdddd 


TfrHO^OCOOOCOO 

CO<NrHO500CDTtCOCO 

rH'^r^dddddd 


I *, 


rt*rHOOr^Oi^OI>00 


tO CQ 00 CO rH ^4 rH l> Oi 

CO<NOOiOOcOOCOC^ 


^ CO 


rHrHrHOOOOOO 


rHrHrHOOOOOO 


"~ ~ j 


^^ i*^ 00 ^^ O3 CO ^5 CO I s * 
CO C^ ^^ O5 t^ 1 * CO lO CO C^ 


COOiCO^OOrHrHOOrH 
COrHOO5t>COlOCOCO 


4) C^ 


rHrHrHOOOOOO 


rHrHrHOOOddd 


|* 


COC<IOOiOOcOOCOC<l 

rHrHrHOOOOOO 


CO C^ ^> O5 t^> CO ^O CO CO 
rHrHrHOOdddd 


o 


CO C^ ^^ Oi 00 CO ^O CO C^ 


a86 88 SS-S 8 


* ^e. 


rHrHrHOOOOOO 


rn'rHr-idddddd 


1 ' " 


CO C^l ^^ Oi t>* CO lO CO C^ 


COrHOOiOOcDtOCOCO 


I 1 * 


rn'r-irn'oddddd 


rHrHrHOOOOOO 


1 A 
I <*= 


rHrHrHOOOddd 


COrHOOit^CO^COC^ 
rHrHrHOOOddd 


*. 


co-^goicgo^^ 


O.O^^^rHO^^O 


r 


r-Jr-^rn'oddddd 


rHrHrHOOdddd 







104 



C. M. TUTTLE AND M. E. RUSSELL [J. S. M. P. E. 



Table II, which is an analysis of the values given in Table I, gives 
the average values obtained by the two methods, the spread in values 
(that is, the differences between the highest and lowest individual 
readings), and the average deviation for a single reading by each 
method. The results for spread in density show that in no case did 
the values of density obtained during the course of the test differ 
by more than 0.02, when read physically. The readings for about 
half the steps differed by only 0.01, and for one step no variation at 

TABLE II 

Motion Picture Negative Material Clear Support 



Step 


Average 
Physical Visual 


Spread 
Physical Visual 


Average 
Physical 


Deviation 
Visual 


2 


.34 


1.33 


0.02 


0.03 


0.004 


0.009 


3 


.26 


1.26 


0.02 


0.03 


0.006 


0.009 


4 


.21 


1.20 


0.01 


0.03 


0.003 


0.010 


5 


.15 


1.14 


0.02 


0.04 


0.006 


0.008 


6 


.08 


1.08 


0.02 


0.04 


0.005 


0.009 


7 


1.01 


1.01 


0.02 


0.03 


0.008 


0.007 


8 


0.94 


0.94 


0.02 


0.03 


0.002 


0.008 


9 


0.87 


0.87 


0.01 


0.03 


0.005 


0.008 


10 


0.79 


0.79 


0.01 


0.03 


0.004 


0.008 


11 


0.70 


0.70 


0.01 


0.03 


0.003 


0.012 


12 


0.63 


0.63 


0.01 


0.04 


0.005 


0.009 


13 


0.55 


0.56 


0.01 


0.03 


0.003 


0.011 


14 


0.50 


0.50 


0.00 


0.04 


0.000 


0.010 


15 


0.42 


0.42 


0.01 


0.04 


0.004 


0.007 


16 


0.37 


0.37 


0.01 


0.02 


0.003 


0.004 


17 


0.32 


0.32 


0.02 


0.02 


0.006 


0.008 


18 


0.28 


0.30 


0.01 


0.03 


0.005 


0.008 


19 


0.27 


0.27 


0.01 


0.02 


0.005 


0.008 


Average 




0.013 


0.031 


0.004 


0.009 



all was found. The same strip, when read by the visual method, 
however, showed a small but definitely larger spread in values for the 
ten readings than that obtained by the physical instrument. Of 
the eighteen steps read, five had a spread of 0.04, ten had a spread of 
0.03, and three had a spread of 0.02. In no case was the spread less 
than 0.02. 

When the results obtained for the average deviation for a single 
reading are examined it will be noted that the results for the physical 
instrument vary from 0.00 to 0.008, with an average for all steps 
of 0.004. When read by the visual method, the average deviation 



Jan., 1937] AN AUTOMATIC RECORDING DENSITOMETER 



105 



for a single reading varies from =*= 0.004 to =*=0.011 with an average 
of ==0.009; i. e., the uncertainty of a single reading made by the 
visual method is at least twice that of a reading made by the physical 
method. 

Since the sensitometric characteristic is usually determined by 
drawing the smooth curve that best represents the composite result 
for fifteen or more points, it will be realized that the uncertainty 
of making an individual visual reading is of no very great importance. 
To illustrate the point, the curves in Figs. 2 and 3 were plotted from 



Or MATERIAL OC- 
IAOC BY THE *4Y 



OCNITOMCTt 



i 



FIG. 2. Curve of material determined from readings made by the 
physical densitometer. 

physical and visual density readings of the same strip. The curves 
are practically identical, although the scattering of the visual points 
about the line is appreciably greater than the scattering of the 
physical density values. 

Thus, for routine sensitometry it is doubtful whether the increase 
of accuracy of the physical method over the visual method is of great 
importance. It may be mentioned, however, that the physical 
instrument has a distinct advantage over the visual method in that 
it is not subject to the likelihood of mistakes in readings or settings 
or to the possible skipping of steps. Moreover, its accuracy is un- 
impaired by the fatigue that affects the quality of results obtained 
by the visual methods. 

There is one case for which the additional accuracy of the physical 
instrument is very important. For some types of sensitometric 



106 



C. M. TUTTLE AND M. E. RUSSELL [J. S. M. P. E. 



work it is necessary to determine with high precision a point on the 
characteristic curve having a certain slope or gradient . If the gradient 
is very low, it usually lies on the toe of the curve, and if the exact 
slope of the curve in that region is not known with high precision, it 
becomes extremely difficult to determine definitely the exposure 
required to produce the gradient in question. 

Table III gives results for a motion picture positive material on a 
clear base similar to those given in Table II for a motion picture 
negative material. It will be noticed that the average results ob- 
tained by both methods for this material are practically identical, 























! 
















CURVE 

DINGS 


OF M/ 
MADE 


ITERIAJ 
BY THE 


- DETE 

VISUA 


PMINE 
- DENS 


D FRON 
ITOME 










' 








BE A 


TER 






^ 


-^ - 
























^ 


^^ 




' 




















jS 


^ 








- 














^ 


S" 


^0 












\ - 












^ 


<< 






















a- 


^_ 


"*"" 










3UPPOP 


T FCK 








- 


























LOG E 


IPOSUR 


E 


















- 



FIG. 3. Curve of material determined from readings made by the 
visual densitometer. 



showing that the difference in diffusion characteristics between 
fine-grained positive and course-grained negative materials has no 
effect upon the results produced by the physical instrument. The 
spreads in density values obtained by both methods are again about 
the same as those obtained for the negative material. Likewise, 
the average deviation for a single reading is the same for both 
methods, as was true in the former case. 

The results discussed so far were from measurements made on 
clear- or gray-base film. Similar measurements were made also on 
motion picture positive material coated on green and red supports. 
A summary of the results obtained for the green-base material is 
given in Table IV. Because of the difficulty of making a visual 
match for a highly colored strip, it is common practice in visual 



Jan., 1937] AN AUTOMATIC RECORDING DENSITOMETER 



107 



densitometry to place a piece of the fixed-out support in the compari- 
son beam so that the color of the two sides of the photometric field 
is the same. With the physical instrument measurements were made 
including the density of the colored strip. It will be noted, therefore, 
that the values obtained by the physical method are somewhat 
higher than those made by the visual method. The density of the 
colored support measured by the physical instrument was found to be 
0.20. When this density is subtracted from the total density for each 
strip measured by the physical instrument, the agreement of results 
with those obtained by the visual method is as good as those ob- 
tained for materials coated on clear or gray supports. 



TABLE HI 



Motion Picture Positive Material Clear Support 



Step 


Average 
Physical Visual 


Spread 
Physical Visual 


Average 
Physical 


Deviation 
Visual 


8 


2.59 


2.60 


0.03 


0.03 


0.007 


0.011 


Q 


2.41 


2.40 


0.01 


0.04 


0.005 


0.012 


10 


2.14 


2.13 


0.02 


0.04 


0.007 


0.011 


11 


1.86 


1.86 


0.02 


0.03 


0.005 


0.009 


12 


1.54 


1.54 


0.01 


0.04 


0.004 


0.010 


13 


1.23 


1.24 


0.01 


0.03 


0.004 


0.009 


14 


0.93 


0.94 


0.01 


0.02 


0.004 


0.005 


15 


0.68 


0.69 


0.01 


0.04 


0.002 


0.016 


16 


0.47 


0.47 


0.01 


0.04 


0.004 


0.007 


17 


0.29 


0.30 


0.01 


0.03 


0.005 


0.012 


18 


0.18 


0.19 


0.01 


0.04 


0.005 


0.010 


19 


0.12 


0.13 


0.00 


0.03 


0.000 


0.008 


20 


0.09 


0.09 


0.00 


0.03 


0.000 


0.005 


Average 




0.012 


0.034 


==0.004 


0.009 



While the spread in density values for the green support obtained 
by the automatic instrument is the same as was noted for the other 
materials, the spread in visually measured values, is noticeably larger. 
The average deviation for a single reading by the physical method 
is exactly the same as for the clear-base materials, but the average 
deviation for readings made visually is distinctly greater (especially 
for the higher densities) than was noted for other materials, in spite 
of the fact that color differences in the field were eliminated. The 
loss of visual accuracy in the reading of colored materials is probably 



108 



C. M. TUTTLE AND M. E. RUSSELL [J. S. M. P. E. 



the result of the decrease in contrast sensitivity of the eye to small 
differences in brightness for light of certain colors as compared with 
white light. A similar effect may be noted in Table V. The aver- 
age deviation for a single reading made by the visual method for a 
red-base material is much greater than for a clear-base material. 

The values in Table V for a motion picture positive material 
coated on a red base are given in the same manner as those in Table 
IV for a green-base material. When allowance is made for the 
density as measured by the physical instrument, the densities ob- 
tained by the visual and physical methods are nearly identical for 
every step. The spread in density and the average deviation ob- 

TABLE IV 

Motion Picture Positive Material Green Support 



Step 


Average 
Physical 
With Minus 
Base Base 


Visual 
Minus 
Base 


Spread 
Physical Visual 


Average 
Physical 


Deviation 
Visual 


8 


2.61 


2.41 


2.44 


0.02 


0.06 


0.006 


0.013 


9 


2.42 


2.22 


2.21 


0.02 


0.04 


0.006 


0.013 


10 


2.18 


1.98 


1.97 


0.02 


0.04 


0.005 


0.012 


11 


1.88 


1.68 


1.67 


0.01 


0.05 


0.005 


0.015 


12 


1.63 


1.43 


1.41 


0.01 


0.03 


0.004 


0.008 


13 


1.32 


1.12 


1.11 


0.01 


0.05 


0.005 


0.010 


14 


1.04 


0.84 


0.84 


0.01 


0.05 


0.005 


0.014 


15 


0.80 


0.60 


0.61 


0.01 


0.07 


0.002 


0.011 


16 


0.62 


0.42 


0.43 


0.00 


0.02 


0.000 


0.012 


17 


0.45 


0.25 


0.26 


0.01 


0.03 


0.003 


0.007 


18 


0.34 


0.14 


0.15 


0.02 


0.02 


0.005 


0.005 


19 


0.25 


0.05 


0.09 


0.01 


0.03 


0.002 


0.009 


20 


0.21 


0.01 


0.04 


0.01 


0.03 


0.004 


0.004 


Average 




0.012 


0.040 


0.004 


0.010 



tained by the physical method are exactly the same for the red-base 
material as for the other materials. Both the spread and the average 
deviation for values obtained by the visual method are much greater 
for the red-base material than for any of the others studied. 

To summarize the results of this comparison it may be said that 
there is no inherent difference in the average results obtained by the 
two methods; the accuracy of the physical instrument is greater in 
making a single reading on any kind of material and very markedly 
greater with colored supports ; and the instrument is essentially free 
from mistakes that may be made by operators of visual instruments. 



Jan., 1937] AN AUTOMATIC RECORDING DENSITOMETER 



109 



DURABILITY AND RUGGEDNESS 

Although the instrument requires remarkably little attention, 
some improvements and changes in the original model are desirable. 
Most frequent difficulties encountered during the first six months 
of operation were: (1) wear in high-speed parts, such as the drive 
shafts and bearings of the motors and speed reducers; (2) sticking 
in the dash-pot of the piston that draws the diaphragm back; (3) 
pitting and burning of the sparking point and its insulator ; (4) dust 
in the optical system. 

The mechanical defects were largely overcome by making spare 
elements for all high-speed parts and modifying the design so that 
these parts may be readily replaced. 



TABLE V 



Motion Picture Positive Material Red Support 



Step 


Phy 
With 
Base 


Average 
sical 
Minus 
Base 


Visual 
Minus 
Base 


Spread 
Physical Visual 


Average 
Physical 


Deviation 
Visual 


9 


2.76 


2.28 


2.31 


0.04 


0.06 


0.011 


0.030 


10 


2.54 


2.06 


2.07 


0.02 


0.04 


0.007 


0.012 


11 


2.29 


1.81 


1.78 


0.03 


0.05 


0.009 


0.011 


12 


1.96 


1.48 


1.48 


0.02 


0.06 


0.005 


0.017 


13 


1.70 


1.22 


1.20 


0.01 


0.07 


0.004 


0.016 


14 


1.42 


0.94 


0.93 


0.01 


0.04 


0.003 


0.010 


15 


1.17 


0.69 


0.68 


0.01 


0.04 


0.005 


0.011 


16 


0.95 


0.47 


0.48 


0.00 


0.04 


0.000 


0.015 


17 


0.78 


0.30 


0.31 


0.02 


0.02 


0.005 


0.007 


18 


0.66 


0.18 


0.17 


0.01 


0.03 


0.003 


0.010 


19 


0.58 


0.10 


0.10 


0.01 


0.04 


0.002 


0.009 


20 


0.54 


0.06 


0.05 


0.00 


0.03 


0.000 


0.008 


Average 




0.015 


0.043 


0.004 


0.013 



The piston that falls into the dash pot has been made loose-fitting 
to prevent sticking. It now acts merely as a counterweight, and the 
distance to which the diaphragm is moved back now depends solely 
upon the time during which the circuit of the magnetic clutch that 
drives the diaphragm is open. The timing is controlled by the 
switch-operating cam or the shaft driven by the over-running clutch. 

A small point not more than ten mils in diameter must be used 
for the spark discharge, as otherwise the position of the spark will 
vary too much. All wires of this diameter will either oxidize or melt 



110 C. M. TUTTLE AND M. E. RUSSELL [J. S. M. P. E. 

unless they are protected by an insulator making sufficiently close 
contact with the metal to cool it. Porcelain and plastic molded 
materials are unsatisfactory because they pit. Carbon deposits from 
the burned paper gather about the metallic point and materially 
increase the sparking area. Frequent cleaning or replacement of 
such points and insulators was necessary. An insulator made from 
a diamond bored with a 10-mil hole with a platinum wire drawn 
through the hole has proved entirely satisfactory. 

The difficulty of dust in the optical system, particularly on the 
small end of the diaphragm opening, has not yet been entirely avoided. 
Efforts have been made to seal the system more thoroughly, but oc- 
casional cleaning is still necessary. Fortunately, errors arising from 
this cause are quickly discovered by periodically reading a check 
strip. 

The photoelectric control system, including the relays, has given 
no trouble. If any appreciable change in cell sensitivity has taken 
place, both comparison and measuring cell must have changed in the 
same manner. It is believed that the opposed potential circuit with 
barrier layer cells used at the null output point is exceedingly reliable. 
The mechanical relay in well over two million operations has never 
caused any trouble. 

Both lamps are run below their rated voltage, and replacements 
are infrequent. Three weeks is the average life of the measuring 
lamp and that of the galvanometer lamp, two months. 

REFERENCES 

1 TUTTLE, C.: "A Recording Physical Densitometer," /. Opt. Soc. Amer., 26 
(July, 1936), No. 7, p. 282. 

2 TUTTLE, C., AND HIATT, B.C.: "A Note on the Measurement of Photographic 
Densities with a Barrier Type of Photocell," J. Soc. Mot. Pict. Eng., XXVI 
(Feb., 1936), No. 2, p. 195. 

3 JONES, L. A. : "An Instrument (Densitometer) for the Measurement of High 
Photographic Densities," J. Opt. Soc. Amer., 7 (March, 1923), No. 3, p. 231. 

DISCUSSION 

MR. KELLOGG: Does the physical densitometer read the identical spot on 
the step tablet each time? If so, how big is the spot? 

MR. RUSSELL: One of the limitations, of course, of any sort of physical 
densitometer is that of getting plenty of light through the sample. Consequently, 
the instrument is designed for just about all of a type lib sensitometer step, which 
is about 8-mm. square. Our visual instruments are of the Martin's polarization 
type, which read areas not quite that large, but very nearly so. 






Jan., 1937] AN AUTOMATIC RECORDING DENSITOMETER 111 

MR. KELLOGG: I wonder, when using the type that measures smaller areas, 
how much variation occurs due to lack of perfect uniformity of the sample area. 
For example, the Capstaff densitometer, I believe, used a hole only about Vi mm - 
across. How much variation to expect would depend upon the density and type 
of the photographic material, but could you give us a general idea what the 
variation is for a spot of that size? 

MR. RUSSELL: The answer to that would depend somewhat, of course, upon 
the character of the strip being read. For motion picture film that has been care- 
fully exposed, processed, and read, the results obtained by the two methods are 
very similar. For strips handled somewhat more carelessly, dust specks and 
things of that sort may be quite serious in the small field of the Capstaff instru- 
ment, causing the results to vary quite noticeably. 



COMMITTEES 

of the 
SOCIETY OF MOTION PICTURE ENGINEERS 

(Correct to December 20, 1936; additional appointments may be made at any 
time during the year as necessity or expediency may require) 



L. W. DAVEE 
A. S. DICKINSON 



ADMISSIONS 

T. E. SHEA, Chairman 
M. W. PALMER 
H. RUBIN 



H. GRIFFIN 
D. E. HYNDMAN 



O. M. GLUNT 
A. C. HARDY 



BOARD OF EDITORS 
J. I. CRABTREE, Chairman 



L. A. JONES 

G. E. MATTHEWS 



W. H. CARSON 
O. O. CECCARINI 



COLOR 

J. A. BALL, Chairman 
C. H. DUNNING 
R. M. EVANS 

A. M. GUNDELFINGER 



H. W. MOYSB 
A. WARMISHAM 



H. GRIFFIN 



CONVENTION 

W. C. KUNZMANN, Chairman 

J. H. KURLANDER 



M. W. PALMER 



H. BUSCH 

A. S. DICKINSON 

G. C. EDWARDS 



EXCHANGE PRACTICE 

T. FAULKNER, Chairman 
A. HIATT 
J. S. MACLEOD 



N. F. OAKLBY 
H. RUBIN 
J. H. SPRAY 



J. E. ABBOTT 
T. ARMAT 



A. N. GOLDSMITH 
A. C. HARDY 



HISTORICAL 

E. THEISEN, Chairman 
G. A. CHAMBERS 
W. CLARK 

HONORARY MEMBERSHIP 

J. G. FRAYNB, Chairman 



G. E. MATTHEWS 
T. RAMSAYE 



H. G. TASKER 
W. E. THEISEN 



112 



COMMITTEES OF THE SOCIETY 



113 



E. HUSB 

K. F. MORGAN 



JOURNAL AWARD 
A. C. HARDY. Chairman 



G. F. RACKETT 

E. A. WlLLIFORD 



J. CRABTRBB 
R. M. EVANS 
E. HUSB 
T. M. INGMAN 



LABORATORY PRACTICE 

D. E. HYNDMAN, Chairman 

M. S. LESHING H. W. MOYSE 



C. L. LOOTBNS 
R. F. MITCHELL 



J. M. NICKOLAUS 
W. A. SCHMIDT 
J. H. SPRAY 



MEMBERSHIP AND SUBSCRIPTION 

E. R. GBIB, Chairman 



Atlanta 

C. D. PORTER 

Boston 

T. C. BARROWS 
J. R. CAMERON 
J. S. CIFRB 

Camden & Philadelphia 
H. BLUMBERG 
J. FRANK, JR. 

Chicago 
B. W. DEPUE 
J. H. GOLDBERG 
S. A. LUKES 
R. F. MITCHELL 



Cleveland 
R. E. FARNHAM 
J. T. FLANNAGAN 
V. A. WELMAN 

Hollywood 
J. O. AALBERG 
L. E. CLARK 
G. H. GIBSON 
C. W. HANDLEY 

E. HUSE 

F. E. JAMES 

G. A. MITCHELL 
P. MOLE 

K. F. MORGAN 
G. F. RACKETT 

Minneapolis 
C. L. GREENE 



New York 

G. C. EDWARDS 

J. J. FINN 

G. P. FOUTE 

H. GRIFFIN 

W. W. HENNESSEY 

R. C. HOLSLAG 

M. D. O'BRIEN 

F. H. RICHARDSON 
H. B. SANTEE 
T. E. SHEA 
J. L. SPENCB 
J. H. SPRAY 

Rochester 

E. K. CARVER 

Washington 
N. GLASSER 

F. J. STORTY 



Australia 
H. C. PARISH 

Austria 

P. R. VON SCHROTT 

China 

R. E. O'BOLGER 



Canada 

F. C. BADGLEY 

C. A. DENTELBBCK 

G. E. PATTON 



England 

W. F. GARLING 

R. G. LlNDBRMAN 
D. McMASTER 

R. TERRANEAU 
S. S. A. WATKINS 



114 



COMMITTEES OF THE SOCIETY 



[J. S. M. P. E. 



France 


India 


Russia 


L. J. DIDIEE 


H. S. MEHTA 


A. F. CHORINE 


L. G. EGROT 


M. L. MISTRY 


E. G. JACHONTOW 


F. H. HOTCHKISS 


M. B. PATEL 




J. MARETTE 




Travelling 




Japan 


E. AUGER 


Germany 


T. NAGASE 


K. BRENKERT 


W. F. BIELICKE 


Y. OSAWA 


W. C. KUNZMANN 


K. NORDEN 




D. McRAE 


Hawaii 


New Zealand 


O. F. NEU 


L. LACHAPELLE 


C. BANKS 


H. H. STRONG 



J. E. ABBOTT 
T. ARMAT 



O. B. DEPUE 



D. P. BEAN 
F. E. CARLSON 
W. B. COOK 
H. A. DEVRY 



MUSEUM 
(Eastern) 

M. E. GILLETTE, Chairman 
H. T. COWLING 
G. E. MATTHEWS 

(Western) 
E. THEISEN, Chairman 

J. A. DUBRAY 



NON-THEATRICAL EQUIPMENT 
R. F. MITCHELL, Chairman 
E. C. FRITTS 
H. GRIFFIN 
R. C. HOLSLAG 



T. RAMSAYE 
E. I. SPONABLE 



A. REEVES 



J. H. KURLANDER 

E. Ross 
A. SHAPIRO 
A. F. VICTOR 



C. N. BATSEL 
L. N. BUSCH 
A. A. COOK 
L. J. J. DIDIEE 



PAPERS 

G. E. MATTHEWS, Chairman 

M. E. GILLETTE H. B. SANTEE 

E. W. KELLOGG T. E. SHEA 

R. F. MITCHELL P. R. VON SCHROTT 



W. A. MUELLER 



I. D. WRATTEN 



J. L CRABTREE 
A. S. DICKINSON 
R. EVANS 



PRESERVATION OF FILM 

J. G. BRADLEY, Chairman 
M. E. GILLETTE 
C. L. GREGORY 



T. RAMSAYE 
V. B. SEASE 
W. A. SCHMIDT 



M. ABRIBAT 
L. N. BUSCH 
A. A. COOK 
R. M. CORBIN 
J. A. DUBRAY 



PROGRESS 

J. G. FRAYNE, Chairman 
R. E. FARNHAM 
E. R. GEIB 
G. E. MATTHEWS 
H. MEYER 
V. E. MILLER 



R. F. MITCHELL 
G. F. RACKETT 
P. R. VON SCHROTT 
S. S. A. WATKINS 
I. D. WRATTEN 



Jan., 19371 



COMMITTEES OF THE SOCIETY 



115 



PROGRESS AWARD 

A. N. GOLDSMITH, Chairman 



M. C. BATSEL 




C. DRBHBR 


J. I. CRABTREE 




J. G. FRAYNE 




PROJECTION PRACTICE 






H. RUBIN, Chairman 




J. O. BAKER 


J. J. FINN 


R. MIEHLING 


T. C. BARROWS 


E. R. GEIB 


E. R. MORIN 


F. E. CAHILL 


A. N. GOLDSMITH 


M. D. O'BRIEN 


J. R. CAMERON 


H. GRIFFIN 


F. H. RICHARDSON 


G. C. EDWARDS 


J. J. HOPKINS 


J. S. WARD 


J. K. ELDERKIN 


C. F. HORSTMAN 


V. WBLMAN 




P. A. McGuiRE 




PROJECTION SCREEN BRIGHTNESS 




C. TUTTLE, Chairman 




A. A. COOK 


W. F. LITTLE 


B. SCHLANGER 


A. C. DOWNES 


O E. MILLER 


A. T. WILLIAMS 


D. E. HYNDMAN 


G. F. RACKETT 


S. K. WOLF 




H. RUBIN 






PUBLICITY 






W. WHITMORE, Chairman 




J. R. CAMERON 


G. E. MATTHEWS 


P. A. McGuiRE 


J. J. FINN 




F. H. RICHARDSON 




SOUND 






P. H. EVANS, Chairman 




M. C. BATSEL 


K. F. MORGAN 


R. O. STROCK 


L. E. CLARK 


O. SANDVIK 


H. G. TASKER 


F. J. GRTGNON 


E. I. SPONABLE 


S. K. WOLF 




STANDARDS 






E. K. CARVER, Chairman 




P. ARNOLD 


P. H. EVANS 


N. F. OAKLEY 


F. C. BADGLEY 


R. E. FARNHAM 


G. F. RACKETT 


M. C. BATSEL 


C. L. FARRAND 


W. B. RAYTON 


L. N. BUSCH 


G. FRIEDL, JR. 


C. N. REIFSTECK 


W. H. CARSON 


H. GRIFFIN 


H. RUBIN 


A. CHORINE 


R. C. HUBBARD 


O. SANDVIK 


A. COTTET 


E. HUSE 


H. B. SANTEE 


L. DE FEO 


C. L. LOOTENS 


J. L. SPENCE 


A. C. DOWNES 


W. A. MACNAIR 


A. G. WISE 


J. A. DUBRAY 


K. F. MORGAN 


I. D. WRATTEN 




T. NAGASB 





116 COMMITTEES OF THE SOCIETY 

STUDIO LIGHTING 
R. E. FARNHAM, Chairman 
W. C. KUNZMANN V. E. MILLER E. C. RICHARDSON 

J. H. KURLANDER G. F. RACKETT F. WALLER 

SECTIONS OF THE SOCIETY 

(Atlantic Coast) 

G. FRIEDL, JR., Chairman 

L. W. DAVEE, Past-Chairman M. C. BATSEL, Manager 

D. E. HYNDMAN, Sec.-Treas. H. GRIFFIN, Manager 

(Mid-West) 

C. H. STONE, Chairman 

R. F. MITCHELL, Past-Chairman O. B. DEPUE, Manager 

S. A. LUKES, Sec.-Treas. B. E. STECHBART, Manager 



SOCIETY ANNOUNCEMENTS 
HOLLYWOOD CONVENTION 



HOTEL ROOSEVELT 
MAY 24-27, 1937 



As announced in previous issues of the JOURNAL, the next Convention of the 
Society will be held at Hollywood, Calif., May 24th-27th, inclusive. Head- 
quarters of the Convention will be the Hotel Roosevelt, where excellent accom- 
modations are assured. A reception suite will be provided for the Ladies' Com- 
mittee, and an excellent program of entertainment will be arranged for the ladies 
who attend the Convention. 

Special hotel rates, guaranteed to SMPE delegates, European plan, will be as 
follows: 

One person, room and bath $3 . 50 

Two persons, double bed and bath 5 . 00 

Two persons, twin beds and bath 6 . 00 

Parlor suite and bath, 1 person 9.00 

Parlor suite and bath, 2 persons 12.00 

Room reservation cards will be mailed to the membership of the Society in the 
near future, and everyone who plans to attend the Convention should return his 
card to the hotel promptly in order to be assured of satisfactory accommodations. 

Special garage rates will be provided for SMPE delegates who will motor to the 
Convention, and arrangements for golfing and other diversions will be announced 
by the Local Arrangements Committee in the near future. 

The dates of the Convention have been chosen in order that delegates may avail 
themselves of the summer tourists' rates, which go into effect May 15th. The 
following table lists the railroad fares and Pullman charges: 

Railroad Fare Pullman 

City (round trip) (one way) 

Washington $120.75 $20.50 

Chicago 86.00 15.75 

Boston 132.80 22.25 

Detroit 98.30 18.00 

New York 126.90 21.75 

Rochester 112.50 19.25 

Cleveland 101.35 18.00 

Philadelphia 122.85 21.00 

Pittsburgh 107.10 18.75 

117 



118 SOCIETY ANNOUNCEMENTS [J. S. M. P. E. 

The railroad fares given above are for round trips, forty-five day limits. Ar- 
rangements can be made with the railroads to take different routes going and 
coming, if so desired, but once the choice is made it must be adhered to, as 
changes in the itinerary may be effected only with considerable difficulty and 
formality. 

New streamlined trains will be operating from Chicago to Los Angeles and 
San Francisco, making the trip to Los Angeles in 39 hours. A special fare is 
levied on these trains. 

Plans for the Convention are already under way. The Papers Committee, 
under the Chairmanship of Mr. G. E. Matthews, and directed by Mr. J. I. Crab- 
tree, Editorial Vice-P resident, are already engaged in soliciting an outstanding 
group of technical papers and presentations. The Officers and Board of Managers 
of the West Coast Section, Mr. K. F. Morgan, Chairman, are collaborating with 
Mr. W. C. Kunzmann, Convention Vice-P resident, in arranging the various facili- 
ties of the Convention. 

The usual Semi-Annual Banquet will be held, and arrangements will be made 
for visits to studios and to other points of interest in and about Hollywood. De- 
tails of the program will be announced later, but in the meantime members are 
urged to make their plans early for attending the Convention, and it is suggested 
that they may perhaps combine their vacation periods with the trip. 



ATLANTIC COAST SECTION 



At a meeting held on December 3rd in the auditorium of the Electrical Associa- 
tion of New York, several reels of film were projected, through the courtesy of 
Mr. E. A. Williford, of National Carbon Company, illustrating the various proc- 
esses applied in making carbons for motion pictures and other purposes. The 
pictures were accompanied by running comments describing the processes and 
their functions in producing the finished carbons. 

Following this presentation, two reels of motion pictures, made at the M-G-M 
Studios of Hollywood, illustrating various cinematographic applications of the 
new polarizing material, Polaroid, were shown. These reels had previously been 
projected at the Rochester Convention. 

As a result of the recent elections, the Officers and Board of Managers of the 
Atlantic Coast Section for the year 1937 will be as follows: 



G. FRIEDL, JR., Chairman 
L. W. DAVEE, Past-Chairman 
D. E. HYNDMAN, Sec.-Treas. 
M. C. BATSEL, Manager 
H. GRIFFIN, Manager 



The Chairman, Past-Chairman, and Secretary-Treasurer are elected for one- 
year terms; of the Managers, the term of Mr. Batsel has yet one year to run; 
that of Mr. Griffin, two years. 



Jan., 1937] SOCIETY ANNOUNCEMENTS 119 

MID-WEST SECTION 

At a meeting held on December 10th, at the manufacturing plant of Motiograph, 
Inc., Chicago, a paper was presented by Mr. E. J. Wienke on "The Adaptability 
of the Motiograph Projector to Sound." Following the paper, several reels of 
film were projected showing third-dimensional effects achieved by the Fleischer 
Studios in producing animated cartoons. These films had previously been pro- 
jected at the Rochester Convention in conjunction with a paper on the subject 
appearing elsewhere in this issue of the JOURNAL. 

As a result of the recent elections, the Officers and Board of Managers of the 
Mid- West Section for the year 1937 will be as follows: 

C. H. STONE, Chairman 

R. F. MITCHELL, Past-Chairman 

S. A. LUKES, Sec.-Treas. 

O. B. DEPUE, Manager 

B. E. STECHBART, Manager 



SOCIETY SUPPLIES 

The following are available from the General Office of the Society, at the prices 
noted. Orders should be accompanied by remittances. 

Aims and Accomplishments. An index of the Transactions from October, 
1916, to December, 1929, containing summaries of all articles, and author and 
classified indexes. One dollar each. 

Journal Index. An index of the JOURNAL from January, 1930, to December, 
1935, containing author and classified indexes. One dollar each. 

SMPE Standards. Reprints of SMPE Standards and Recommended Practice. 
Twenty-five cents each. 

Membership Certificates. Engrossed, for framing, containing member's name, 
grade of membership, and date of admission. One dollar each. 

Lapel Buttons. The insignia of the Society, gold filled, with safety screw back. 
One dollar each. 

Journal Binders. Black fabrikoid binders, lettered in gold, holding a year's 
issue of the JOURNAL. One dollar each. Member's name and the volume 
number lettered in gold upon the backbone at an additional charge of fifty cents 
each. 



STANDARD S. M. P. E. 

VISUAL AND SOUND TEST REELS 

Prepared under the Supervision 

OF THE 
PROJECTION PRACTICE COMMITTEE 

OF THE 
SOCIETY OF MOTION PICTURE ENGINEERS 



Two reels, each approximately 500 feet long, of specially pre- 
pared film, designed to be used as a precision instrument in 
theaters, review rooms, exchanges, laboratories, and the like 
for testing the performance of projectors. The visual section 
includes special targets with the aid of which travel-ghost, 
lens aberration, definition, and film weave may be detected 
and corrected. The sound section includes recordings of 
various kinds of music and voice, in addition to constant- 
frequency, constant-amplitude recordings which may be used 
for testing the quality of reproduction, the frequency range 
of the reproducer, the presence of flutter and 60 -cycle or 96- 
cycle modulation, and the adjustment of the sound-track. 
Reels sold complete only (no short sections). 

PRICE $37.50 FOR EACH SECTION, 
INCLUDING INSTRUCTIONS 

(Shipped to any point in the United States) 

Address the 

SOCIETY OF MOTION PICTURE ENGINEERS 

HOTEL PENNSYLVANIA 
NEW YORK, N. Y. 



JOURNAL 

OF THE SOCIETY OF 

MOTION PICTURE ENGINEERS 

Volume XXVIII FEBRUARY, 1937 Number 2 



CONTENTS 

Page 

The Projection of Lenticular Color-Films 

J. G. CAPSTAFF, O. E. MILLER, AND L. S. WILDER 123 

Trick and Process Cinematography J. A. NORLING 136 

Slide-Rule Sketches of Hollywood H. G. TASKER 158 

Effect of Light-Source Size with 16-Mm. Optical Systems 

G. MILI 164 

Medical Motion Pictures in Color 

R. P. SCHWARTZ AND H. B. TUTTLE 173 

New Method for the Dry Hypersensitization of Photographic 
Emulsions F. DERSCH AND H. DURR 178 

Report of the Membership and Subscription Committee 188 

New Motion Picture Apparatus 

A New High-Quality Portable Film-Recording System 

F. L. HOPPER, E. C. MANDERFELD, AND R. R. SCOVILLE 191 

A New High-Quality Film Reproducer J. C. DAVIDSON 202 

Recent Developments in High- Intensity Arc Spotlamps for 

Motion Picture Production E. C. RICHARDSON 206 

The Schwarzkopf Method of Identifying Criminals 

J. FRANK, JR. 212 

Committees of the Society 218 

Spring, 1937, Convention; Hollywood, Calif., May 24th-28th, 
inclusive 223 

Society Announcements 228 



JOURNAL 

OF THE SOCIETY OF 

MOTION PICTURE ENGINEERS 



SYLVAN HARRIS, EDITOR 

Board of Editors 
J. I. CRABTREE, Chairman 

A. N. GOLDSMITH L. A. JONES H. G. KNOX 

A. C. HARDY E. W. KELLOGG T. E. SHEA 



Subscription to non-members, $8.00 per annum; to members, $5.00 per annum, 
included in their annual membership dues; single copies, $1.00. A discount 
on subscriptions or single copies of 15 per cent is allowed to accredited agencies. 
Order from the Society of Motion Picture Engineers, Inc., 20th and Northampton 
Sts., Easton, Pa., or Hotel Pennsylvania, New York, N. Y. 
Published monthly at Easton, Pa., by the Society of Motion Picture Engineers. 

Publication Office, 20th & Northampton Sts., Easton, Pa. 
General and Editorial Office, Hotel Pennsylvania, New York, N. Y. 

West-Coast Office, Suite 226, Equitable Bldg., Hollywood, Calif. 
Entered as second class matter January 15, 1930, at the Post Office at Easton, 
Pa., under the Act of March 3, 1879. Copyrighted, 1937, by the Society of 
Motion Picture Engineers, Inc. 

Papers appearing in this Journal may be reprinted, abstracted, or abridged 
provided credit is given to the Journal of the Society of Motion Picture Engineers 
and to the author, or authors, of the papers in question. Exact reference as to 
the volume, number, and page of the Journal must be given. The Society is 
not responsible for statements made by authors. 



OFFICERS OF THE SOCIETY 

President: S. K. WOLF, 250 W. 57th St., New York, N. Y. 
Past-President: H. G. TASKER, Universal City, Calif. 

Executive Vice-President, G. F. RACKETT, 823 N. Seward St., Hollywood, Calif. 
Engineering Vice-President: L. A. JONES, Kodak Park, Rochester, N. Y. 
Editorial Vice-President: J. I. CRABTREE, Kodak Park, Rochester, N. Y. 
Financial Vice-President: O. M. GLUNT, 463 West St., New York, N. Y. 
Convention Vice-President: W. C. KUNZMANN, Box 6087, Cleveland, Ohio. 
Secretary: J. FRANK, JR., 90 Gold St., New York, N. Y. 
Treasurer: L. W. DAVEE, 250 W. 57th St., New York, N. Y. 

GOVERNORS 

M. C. BATSEL, Front and Market Sts., Camden, N. J. 

A. S. DICKINSON, 28 W. 44th St., New York, N. Y. 

G. FRIEDL, JR., 250 W. 57th St., New York, N. Y. 

A. N. GOLDSMITH, 444 Madison Ave., New York, N. Y. 

H. GRIFFIN, 90 Gold St., New York, N. Y. 

A. C. HARDY, Massachusetts Institute of Technology, Cambridge, Mass. 

K. F. MORGAN, 7046 Hollywood Blvd., Los Angeles, Calif. 

C. H. STONE, 205 W. Wacker Drive, Chicago, 111. 



See p. 218 for Technical Committees 



THE PROJECTION OF LENTICULAR COLOR-FILMS* 
J. G. CAPSTAFF, O. E. MILLER, AND L. S. WILDER** 



Summary. In the projection of lenticular color-films a large portion of the incident 
light is lost by absorption in the tricolor filters . To determine the feasibility of satis- 
factorily showing these films in large theaters, an experimental projector was set up 
embodying the few simple changes in standard theater equipment that were necessary 
to obtain the required large increase in screen illumination. 

Successful demonstrations with the apparatus at Loew's Rochester Theater at 
Rochester and the Center Theater at New York have proved that it is quite possible to 
secure enough screen brightness to give a satisfactory showing of the lenticular films 
in the majority of theaters. 

The principal changes made in the standard projection apparatus in order to ob- 
tain the greatly increased illumination were as follows: 

(1) Increased Relative Aperture. By substituting an f/ 1.6 projection lens for 
the f/2.4 lens commonly used, and by increasing the working relative aperture of 
the 65-ampere high-intensity reflector arc so as to take full advantage of the increased 
aperture of the projection lens, it was possible to get 2.25 times the screen illumina- 
tion obtained with the regular equipment. 

(2) Reduction of Shutter Loss. A further increase was obtained by the use of a 
quicker pull-down and a corresponding reduction in the angle of the shutter blades; 
this may not, however, be feasible in practice. 

(3) Increased Filter Transmission. As a result of numerous practical tests it 
was found to be possible to increase the transmission of the tricolor projection filters by 
33 per cent, without undue loss of color values. 

(4) Lower Print Density. The excellent tone reproduction obtained in the 
process, together with a modification of the optics of the lenticular film, makes pos- 
sible a substantial lowering of the print density. The resultant increase in the 
brightness of the projected image amounts to some 25 per cent. 

The large increase in the radiant energy directed upon the film has made it necessary 
to employ a heat filter in the condenser system. 

Refinements in the present system are expected to produce additional small 
increases in illumination, and it is believed to be possible to develop other special 
equipment to take adequate care of the few (special) cases where it is necessary to 
project upon an unusually large screen. 

The lenticular film color process, in common with other additive 
color processes, involves a large loss of light by absorption in the 
color filters necessarily used in the projection system. Therefore, 

* Received October 12, 1936; presented at the Fall, 1936, Meeting at Roches- 
ter, N. Y. Communication No. 605 from the Kodak Research Laboratories. 

** Eastman Kodak Co., Rochester, N. Y. 

123 



124 



CAPSTAFF, MILLER, AND WILDER [J. S. M. P. E. 



it requires so much more illumination than is needed for projecting 
black-and-white pictures that it was believed until recently by 
many persons in the industry that it was impossible to show these 
pictures properly even in the average theater, not to mention the 
de luxe houses having screens from 25 to 35 feet in width. To il- 
lustrate the seriousness of the problem, it was estimated that about 
ten times the normal amount of light would be needed. The color 
filters used for projection during the earlier experimental work had a 
transmission of only 12 or 13 per cent, and the intensity was further 
reduced by the lenticular surface of the film support. The Kodak 
Research Laboratories recently undertook to make a systematic 



COLLECTOR ELEMENT 



PROJECTION LENS 



COLOR FILTER 




FIG. 1. Diagram of projection optical system for lenticular color-film. 

investigation of the possibilities of lenticular film projection and to 
give an actual demonstration in a de luxe theater. 

A preliminary survey of the problem indicated quite a number 
of possible ways in which the screen illumination could be increased. 
vSome of these, which were temporarily laid aside for practical reasons, 
will not be mentioned except in the concluding remarks. With a 
desire to limit the investigation to the use of already existing pro- 
jection equipment with only minor alterations, the work was pursued 
along the following lines : 

(1} Reduction of the absorption loss in the color niters. 

(2) Modification of the optical system to increase its relative aperture. 

(.V) Recovery of part of the light lost because of the shutter. 

(4) Reduction of the density of the prints. 

( J) Improvement in the operating conditions of the illuminating system. 

EXPERIMENTAL 

Filters 

Since the greater part of the light is lost due to absorption in the 
color filters, the problem of screen brightness becomes progressively 



Feb., 1937] PROJECTION OF COLOR-FILMS 125 

easier as the filter transmission is increased. After a certain point, 
however, the colors of the projected picture begin to lose saturation, 
and appear "washed out." The color reproduced upon the screen 
can be of no higher degree of purity than that of the projection filters. 
As the red filter is made lighter, it soon begins to transmit yellow, and 
becomes an orange-red. With such a filter a good red can not be 
represented properly upon the screen. After considerable experi- 
mental work with dyes, and a number of observations with filters of 
different densities, a standard filter was finally adopted that was 
thought to have the highest transmission it was possible to get with- 
out too noticeable loss in color saturation. The transmission of 
this filter, when used with the high-intensity arc system to be de- 
scribed later, was 22 per cent. This multiplied by the 80 per cent 
transmission of the lenticular film support gives an overall transmis- 
sion of 17.6 per cent. Therefore, the factor by which the normal 
illumination needs to be increased is 5.8 times. 

Optical System 

Fundamental Conditions. As shown in Fig. 1, the essential elements 
of a projection system suitable for lenticular color-films are: light- 
source, collective element, collimator lens, film-gate, projection lens, 
and color filter. A detailed discussion of the optical relations involved 
in the use of lenticular films is not within the scope of this paper, and 
therefore a mere statement is made of the necessary conditions to be 
observed in practice: 

(1) The light-source must be imaged at the film-gate. 

(2} The collecting element must be imaged at the color filter. 

(5) It is essential that all elements be centered upon the optical axis. 

(4) The color filter must be located at the front focus of the projection lens. 

It will be obvious that the first three of these are the identical 
conditions for optimal screen brightness and uniformity, even in 
black-and-white projection. The fourth condition comes about as 
a result of a particular optical property of the lenticular color-film 
itself, and is dependent upon the optical arrangement used in print- 
ing. 

Projection Lenses. The greatest single gain in illumination prom- 
ised to come from increasing the relative aperture beyond the //2.5 
systems commonly used. In view of the successful use, in the 16-mm. 
field, of lenses having relative apertures of //1. 6 or better, it was 



126 



CAPSTAFF, MILLER, AND WILDER [J. S. M. P. E. 



thought that it should be possible to set up a 35-mm. system that 
would be equally efficient. Two //1. 6 lenses were obtained having 
focal lengths of 120 and 160 millimeters. Except for the somewhat 
inferior definition of one of them, these lenses were entirely satisfac- 
tory for the purpose. On account of the much larger diameter of the 
lens barrel, it was necessary to make a new lens mount for the Simplex 
projector. 

Illuminating Systems. Before the increased relative aperture could 
be fully realized, it was necessary to modify the existing illuminating 
system so as to fill an angle of //1. 6 and at the same time to fulfill the 




FIG. 2. Scale drawing of experimental projection system. 

conditions necessary for use with lenticular films. The lamp selected 
for the first experiments was the Peerless "Magnarc," * which appeared 
to be a good example of a high-efficiency reflector system. After a 
number of optical arrangements had been tried, using reflectors of 
various focal lengths, it was apparent that the only change necessary 
was the addition of an inexpensive condenser lens at the front of the 
lamp house. To avoid breakage due to the extreme heat, this lens 
was made of Pyrex. The complete optical arrangement as it was 
finally used is shown in Fig. 2, which is drawn approximately to 
scale. The regular "Magnarc" reflector is 14 inches in diameter, and 



* Other lamps on the market similar to this one should be equally suitable. 



Feb., 1937] PROJECTION OF COLOR-FILMS 127 

5 l /4 inches from the arc crater. The piano surface of the auxiliary 
condenser is 28 inches from the center of the reflector, and 5 J /4 inches 
from the film-gate. This condenser is 4 7 /i 6 inches in diameter and 
15 inches in focal length. The addition of this condenser to the 
"Magnarc" brings the image of the reflector into the plane of the 
three-color projection filter. The filter is located near the front focal 
plane of the projection lens, a necessary condition for lenticular film 
projection. In order to allow the larger cone of illumination from 
the modified illuminating system clear access to the film-gate, it 
became necessary to enlarge the apertures in the shutter housing 
and in the masks back of the aperture plate on the Simplex pro- 
jector. When the full //1. 6 relative aperture is filled, there should 
be 2.31 times the screen brightness that is obtained with a corre- 
sponding system of relative aperture //2. 5. The actual screen bright- 
ness obtained with this system was slightly less due to mild imper- 
fections in the quality of the reflector. Certain dark zones appear 
upon the reflector surface when viewed from the film-gate. This 
modified "Magnarc" system was used for a great part of the ex- 
perimental work and for the demonstrations that are to be mentioned 
presently. 

Magnification of the Arc Crater. It will perhaps be contended that 
the increase in the relative aperture attained in this way is at the 
expense of the crater magnification at the film-gate, and that the 
uniformity in screen brightness will be unsatisfactory. Of course, 
the crater of the high-intensity arc is not uniform in brightness, being 
brighter at the center than at the border. For this reason, and also 
in order to provide some tolerance in the position of the arc, present 
illuminating systems are made to have a higher magnification than 
would be necessary just to fill the aperture. However, when the 
lenticular color-films are projected with the above-described system, 
the corners of the picture do not appear to be more poorly illuminated 
than is the case with the average black-and-white system. The 
reason for this lies in a particular requirement of the camera and 
projector lenses used in the lenticular film process. The lenses used 
in black-and-white work, both in the camera and, to a somewhat 
less extent, in the projector, cause a falling off in the marginal il- 
lumination due to the fact that the lens aperture can not be completely 
filled for oblique angles. 1 With some of the camera lenses ordinarily 
used in black-and-white work, this becomes so bad that the corner 
illumination falls nearly to zero, and results in a print having higher 



12 g CAPSTAFF, MILLER, AND WILDER [J. S. M. P. E. 

density at the corners than at the center of the picture. When this 
print is projected, the additional density at the corners adds consider- 
ably to the deficiency of corner illumination already present in the 
projection system. 

This property of the lenses becomes objectionable in the lenticular 
color-film process, but for a different reason, as seen in Fig. 3, which 
shows different views of the lens and color filters as they would 
appear when viewed from different points of the screen. Dispro- 
portionate areas of the color segments are illuminated for different 





VIEW FROM SIDE 
AXIAL VIEW OF SCREEN 





VIEW FROM TOP VIEW FROM BOTTOM 

OF SCREEN OF SCREE.N 

FIG. 3. Projection lens and color filters seen 
from different points of the screen showing cut- 
off of the filter zones. 

positions around the margin of the screen, a condition that leads to 
an uneven distribution of color on the screen and can not be tolerated. 
Thus, because of the choice of lenses that this makes necessary, one 
can afford to use a lower magnification of the crater. However, it 
may be desirable to have a slightly larger crater image than that used 
in the present system, which could be accomplished by substituting 
a 9- or 10-mm. carbon for the 8-mm. one now used. 

The Heat Problem. Considering that there are already reports from 
theaters when using improved black-and-white equipment of too 
much heat at the picture aperture, it was not surprising to find in the 
preliminary trials with this more efficient optical system that the 



Feb., 1937] PROJECTION OF COLOR-FILMS 129 

film was badly damaged by the terrific heat. Attempts to cool the 
film by a jet of compressed air were insufficient. Clearly some sort 
of heat filter had to be used. Previous experience with water cells 
did not favor their use in the theater projection booth, so heat-ab- 
sorbing glass was tried. Used in a single sheet, it broke repeatedly 
even though it was of the heat-resisting type. Cutting the glass 
into 3 /4-inch strips and mounting the strips side by side prevented 
breakage, but it was found that the glass would soon melt unless 
subjected to a current of air. Since too much color in the glass 
would have been objectionable, it was necessary to use a density only 
just sufficient to reduce the heat to a safe value. The filter finally 
adopted was in the form of several 3 / 4 -inch strips of Corning Extra- 
Light Aklo, 2 millimeters thick, held loosely side by side in a rectangular 
metal frame, and cooled by a gentle current of air from a small furnace 
blower. The location of the filter in the optical system must be 
such that the edges of the glass strips are not visible upon the screen. 
In the present instance, the glass was mounted upon the front of the 
shutter housing at a distance of approximately 3 3 / 8 inches from the 
film-gate. No trace of the edges of the strips has ever been noticed 
upon the screen. The air was directed upon both sides of the glass 
by appropriate baffles With this filter, which transmits only 25 or 
30 per cent of the total heat energy, the heat at the aperture is actually 
less than that occurring with some of the better projection lamps 
now in use. The familiar "biscuit" appearance of projected prints 
is entirely lacking. Part of the air from the blower is directed upon 
the film-gate, which gives slight additional cooling to the film and to 
the metal parts around the aperture. 

A Relay Condenser System. To see what could be done with the 
120-ampere, high-intensity arc used with a condenser system, a Hall & 
Connelly lamp was set up with a set of 7-inch condensers and a relay 
system. In a relay system full advantage can be taken of the entire 
crater surface because it is not imaged at the aperture. Furthermore, 
advantage can be taken of the fact that the entire crater area emits 
red light of practically uniform intensity. Since in color work, the 
limiting color seems to be red, use can be made of the entire crater 
surface. The measurements of screen brightness made with this 
set-up show that it is possible to get equally as bright a screen with 
the "Magnarc" system, and it becomes somewhat easier to maintain 
the screen uniformity. Therefore, where there is sufficient space in 
the projection room to accommodate the increased length of a relay 



130 CAPSTAFF, MILLER, AND WILDER [J. S. M. P. E. 

system, this type of lamp would serve very well. The remarks about 
to be made about adjustment and operation of the optical system 
apply equally well to condenser systems and reflector systems. 

Adjustment and Operation. A great number of observations were 
made with the best types of black-and-white illuminating systems at 
present in use, in order to determine, if possible, what effect the 
operating conditions and the adjustment and alignment of the 
optical system had upon screen brightness. Based upon these 
observations, it is believed probable that the average theater pro- 
jection machine often does not deliver much more than half the 
screen illumination it is capable of delivering. Losses occur in many 
ways: accumulation of dirt upon the screen lowers the reflecting 
power; the reflector or condenser surface facing the arc becomes 
clouded with smoke, pitted by flying particles, and has to be cleaned 
constantly in order to preserve the light transmission. Because of 
the imperfections in the commercial mirrors and condensers, the 
screen uniformity is not at its best when the system is adjusted to 
provide the maximum of screen brightness. 1 The projectionist, 
therefore, has to sacrifice a considerable amount of screen brightness 
in order to improve the uniformity. Errors in centering condenser 
systems can be responsible for appreciable losses of illumination. 
Some projection lenses are in use that have a lower transmission than 
is desirable. Carbon arcs are somewhat erratic in behavior. The 
crater sometimes burns unevenly, and the crater brightness varies 
from time to time. Substantial improvement could be made in all 
these operating conditions. Possibly new equipment would have 
to be designed in order to free the projectionist from the necessity of 
constantly attending to the adjustments of the various manual con- 
trols found upon the present lamps. If the arc operation could be 
sufficiently stabilized, and the arc crater accurately held to the optical 
axis, the entire system could be set up and adjusted once for all, and 
the projectionist would then be required to make only the single 
adjustment of keeping the arc crater in the correct position along the 
optical axis. There is no reason, furthermore, why an arrangement 
using photoelectric cells could not be devised that would make even 
this adjustment automatically. 

Reduction of the Shutter Loss 

Since 50 per cent or more of the incident light is lost at the shutter, 
it seemed worth while to attempt to recover some of this loss by speed- 



Feb., 1937] PROJECTION OF COLOR-FILMS 131 

ing up the pull-down movement, and using shutter blades of the 
narrowest possible angle. No originality is claimed for the method 
used. Inside the housing of the Geneva pull-down mechanism used 
on all Simplex machines there is a pair of small spur gears, through 
which the intermittent assembly is driven. By substituting a pair 
of elliptical gears, the intermittent movement was accelerated so 
that the pull-down period occurred in 52 degrees of the cycle instead 
of the usual 90 degrees. Using this in combination with a 45-degree 
covering blade and a 30-degree flicker blade, a gain of 59 per cent 
was made in screen illumination. However, it was thought that 
this was too severe for the film, and a second pair of elliptical gears was 
prepared that gave a more moderate acceleration to the pull-down 
mechanism, and accomplished the movement of the film in a 68- 
degree interval. Using with this a covering blade of 60 degrees and 
a flicker blade of 40 degrees, a gain of 44 per cent was realized. How- 
ever, unless there are certain changes made in projector design that 
will compensate by reducing the stresses occurring during the pull- 
down operation, it is doubtful whether application of even this mild 
degree of acceleration to the Geneva movement is practicable. The 
Powers movement, however, because of the smooth acceleration, 
offers possibilities for a quicker pull-down. 

The proper size for the shutter blades was arrived at empirically 
by progressively increasing the width until there was no noticeable 
flicker or travel-ghost upon the screen at the ordinary brightness 
level. Advantage was taken in these experiments of the fact that 
the perceptibility of both flicker and travel is less as one proceeds to 
lower levels of illumination. If it should later be found necessary to 
increase the shutter blade slightly, it would represent a loss of only 
a few per cent. A further discussion of the subject of projector 
mechanisms is believed outside the scope of the present paper. Al- 
though the work done so far must be regarded as merely preliminary, 
there seems to be ample ground for believing that more can be done in 
a practical way to recover a considerable part of the light lost at the 
shutter. In this connection, moving the shutter to a position very 
near the film plane so as to effect quicker cut-off of the light-beam 
would be a worth while step. However, in the small neighborhood 
theaters, probably no change in projector mechanism would be 
needed in order to get sufficient light. 

Print Density. Another loss of light occurring in the ordinary pro- 
jector is caused by the minimum photographic density allowable in 



13 2 CAPSTAFF, MILLER, AND WILDER [J. S. M. P. E. 

making the print. Because of the excellent tone reproduction at- 
tained with the lenticular process, it is possible to make the print 
density lower than that of a corresponding black-and-white print by 
approximately 0.10. This gives a 25 per cent increase in picture 
brightness. 

Summary of the Gains Made. It was pointed out above in connec- 
tion with the filters that the maximum filter transmission, combined 
with that of the lenticular support, was in the neighborhood of 17.6 
per cent, which corresponds to a factor of 5.8. This is the factor by 
which the screen brightness must be increased in order to equal that 
of corresponding black-and-white projection. The gains made and 
discussed above may be summarized as in Table I : 

TABLE I 

Factor 

(1) Increased relative aperture (//2.5 to //1. 6) 2 . 3lx 

(2) Reduction of shutter loss 60 to 40 shutter, 68 pull-down 1 .44* 

(3) Lower print density by 0.10 1 . 25x 

Product of all the above gains 4 . 32x 

This is somewhat short of the required 5.8, which is necessary to 
balance the filter loss. In addition to these gains, the authors are 
of the opinion that the screen illumination can be doubled if sufficient 
improvement can be made in the operating conditions of the arc and 
the optical system. The product of this and all the gain factors 
given above leaves ample margin for the projectionist in operat- 
ing the projector when compared to the loss factor of 5.8 men- 
tioned above. 

DEMONSTRATIONS 

The complete experimental projector was used to give two demon- 
strations in the Loew's Rochester Theater in April, 1936. On 
both occasions the 52-degree accelerated pull-down was used. After 
the 6X-degree pull-down was substituted, the machine was used to 
give a demonstration in the Center Theater in Radio City on July 9, 
1936, before some 200 guests. Many of the audience commented 
upon the show, but no one expressed any feeling that there was a 
lack of screen brightness. Some said they actually believed the 
screen brightness was greater than necessary. 

Although many measurements of the screen illumination were 
made throughout all these experiments, a simple statement of the 
values in foot-candles attained would have little meaning in view 






Feb., 1937] PROJECTION OF COLOR-FILMS 133 

of the conflicting reports already published both as to the screen 
brightness actually prevailing in theaters and as to the actual level 
of screen brightness that is to be desired. To give some indication, 
however, of the amount of light obtained on the screen of the Center 
Theater, the value measured with a Weston illumination meter, 
model 603, without the color filters or lenticular film, but with the 
shutter running, was 33 foot-candles at the center of the screen. The 
screen picture was 22 feet wide, and the projection angle was ap- 
proximately 28 degrees. If the heat-absorbing glass filter had been 
removed, the value would have been more than 40 foot-candles. 

FURTHER POSSIBILITIES 

Of course, every precaution was taken in both the demonstrations 
to assure optimal operating conditions. It is probably too much to 
hope that optimal conditions could be thus maintained at all times. 
With this in mind, other possibilities will now be discussed, by means 
of which still more light might be obtained. If the regular high- 
intensity carbons were used, instead of the Suprex carbons, in con- 
nection with a reflection type of lamp of most efficient design, there 
would be an increase due to the higher intrinsic brightness attained 
with the regular high-intensity carbons. The possibilities that a 
new type of arc source will be developed having still higher intrinsic 
brightness can not be excluded. In this connection, carbon manu- 
facturers express the belief that developmental work now in progress 
will produce a carbon that, with the proper optical system and lamp 
mechanism, will give the desired intensity, color, and uniformity of 
light, and, at the same time, keep the energy input into the arc within 
reasonable limits. There are some improvements yet to be made in 
the present experimental optical system that will make it possible 
to eliminate some of the glass-air reflection losses. A desirable 
further improvement in the optical quality of commercial reflectors 
would reduce losses arising from the imperfect formation of the 
crater image at the film-gate. The belief has already been expressed 
that improvement in projector design could be made that would 
further reduce the shutter loss. Another consideration is the possi- 
bility of a slight reduction of the screen size. Even for black-and- 
white projection, a reduction of screen size is being advocated by 
some in the industry. It is difficult to find any objection to doing 
this, since, with the present sizes of screens there is always a large 
block of seats near the front of the theater that the patrons avoid 



!34 CAPSTAFF, MILLER, AND WILDER [J. S. M. P. E. 

because of the discomforts of so large a viewing angle. There would 
seem to be no loss of desirable seating space by making conditions 
more comfortable for those in the front even at the expense of some 
loss in the rear of the house. Since the screen brightness would vary 
inversely as the square of the screen width, a considerable gain in 
illumination ought to be made possible by only a moderate reduction 
of screen size. The use of the ordinary specular screens would, of 
course, be limited to the long narrow houses, in which the seats are 
distributed within an angle of some 20 degrees. The design of 
equipment to take care of the few large houses having exceptionally 
large screens must be considered as a separate problem. 



CONCLUSION 



Although not all possibilities have been utilized in this preliminary 
investigation of the problem, it is seen from the foregoing experi- 
ments that lenticular color-films can be projected satisfactorily in 
the average theater without the necessity of making major altera- 
tions in the present equipment. 



REFERENCE 

1 COOK, A. A.: "A Review of Projector and Screen Characteristics and Their 
Effects upon Screen Brightness," /. Soc. Mot. Pict. Eng., XXVI (May, 1936), 
No. 5, p. 522. 

DISCUSSION 

MR. RICHARDSON: Is it not possible, by means of the additional lens, to 
parallel the light-beam between the aperture and the projection lens, and thus 
have more uniform screen illumination? 

MR. MILLER: Arc lights do not radiate with equal intensity in all directions. 
Something could probably be done by using a shorter focal length reflector. 
Unfortunately none were available at the time, and we wanted to incorporate a 
minimum number of changes in the lamp. 

MR. TASKER: The review room screens at Universal Studio are measured 
daily, in view of the fact that they vary considerably. We find that it is 
possible by daily readjustment of the arc to get such results as 16 foot-lamberts 
at the center, 14Va to 15 at the edges, measured horizontally across top and 
bottom. Within a day's time the lamp is out of adjustment, and may be off as 
much as 15 at the center, 15 on one side, and 11 on the other. 

MR. KURLANDER: A ratio of I 1 /* to 1 from the center to the corners is con- 
sidered excellent. That is about the limit one can get and still maintain good 
efficiency. It is comparatively easy to get evenness if one wants to sacrifice in- 
tensity. 

For 16 years or more we have been trying to get even a 10 per cent increase in 
screen illumination. Mr. Miller has just told us how he obtained a 400 per cent 



Feb., 1937] PROJECTION OF COLOR-FILMS 135 

increase. In view of some of the methods used to attain this 400 per cent increase, 
I should like to see it analyzed a bit more. The mechanical efficiency of the pro- 
jector, over a period of 16 years, has been increased from about 4 per cent to 
8 per cent a 100 per cent increase over a period of 16 years. Here, over night, 
we have another 400 per cent. 

MR. MILLER: We got 2.3 times as much light by increasing the relative aper- 
ture. It was a few per cent less than that when the aperture was increased from 
y/2.5, which is about what is used on the average now, to //1. 6. I purposely 
omitted discussing the optical principles, particularly because they were so 
well treated by A. A. Cook in the JOURNAL of last May. 

MR. McAuLEY : What is the distance from the vertex of the mirror to the aper- 
ture? How much did you move up the projector? 

MR. MILLER: The projector was moved up only very slightly. The distance 
trom the vertex of the reflector to the lens was 28 inches; from there to the 
aperture was 5 l /4 inches, making a total of SS 1 /^ I believe a distance of 34 inches 
is recommended by the lamp manufacturers. 

MR. McAuLEY: Was it necessary to change the reflector? 

MR. MILLER: No. 

MR. McAuLEY: In order to get the increase of speed? It would seem that 
you would hardly fill the lens at a distance of 33 inches. 

MR. MILLER: If the light-rays are drawn backward after passing through the 
auxiliary condenser, it seems from the aperture as if the mirror were bigger. 

MR. BRENKERT: Did you say that with the same mirror that was furnished 
with the lamp, and by adding the collimating lens, you filled an //1. 6 lens? 

MR. MILLER: Yes. 

MR. BRENKERT: And obtained more light upon the screen than by using an 



MR. MILLER: Yes. 

MR. BRENKERT: How did you get more light out of the arc, through the aper- 
ture and trom the mirror, without changing any mirror specifications? 

MR. MILLER: The fact is that the spot upon the aperture plate in the regular 
machine is much larger than the aperture itself, due to imperfections in the 
optical system, and also to the fact that the lamp manufacturer is desirous of 
giving a broad tolerance to the projectionist for keeping his arc focused properly. 
The magnification of the lamp without the condenser was six times. 

MR. BRENKERT: You reduced the size of the spot by means of the auxiliary 
condenser, and that is your sole mention of greater angle and more light. 

MR. MILLER: Yes, but we could use a larger carbon and a little more current, 
and thus increase the size of the spot. We use an 8-mm. carbon; but I think a 
10-mm. would be better. 

MR. BRENKERT: What became of the illumination at the corners? 

MR. MILLER: If a black-and-white film was projected, the corner illumination 
was not good; if a lenticular film, the corners were no worse than for the black- 
and-white, with normal projection as occurring in theaters. 



TRICK AND PROCESS CINEMATOGRAPHY 5 

J. A. NORLING** 



Summary. Process photography, which is the broad classification given to all 
branches of special and trick cinematography, plays an important part in making 
today's motion picture. Many articles have appeared relating to this subject, but, 
unfortunately, most of them have been devoted only to a discussion of the importance 
of this branch of photography, and very few writers have divulged any of the details 
of the methods employed. This paper sets forth in general the underlying procedure 
in the various branches of the art, and treats many phases thereof in sufficient detail to 
be fully informative. 

The branches of process photography disclosed include: transitional effects, 
such as dissolves and wipes; matte shots; simple and intricate multiple exposures; 
composites and montages; animated titles and presentation effects; combined draw- 
ing and actual photography; optical trick printers and cameras; miniature projec- 
tion background process; problems in making dupe negatives by projection, dodging, 
etc. Important steps are described and illustrated, and special devices are shown and 
their essential Junctions and operation described. 



Trick cinematography, as exemplified by double exposures, minia- 
ture settings, and so-called "glass" shots, have been employed to a 
considerable extent since the early days of the motion picture. The 
pioneer workers created some excellent effects, considering their 
facilities. They were handicapped in several ways. Early cameras 
did not provide sufficiently accurate registration to avoid movement 
between individual exposures comprising a composite scene. In 
addition, the photographic emulsions then available were unsuitable 
for duplicating. The first limitation, that of accurate registration, 
has been overcome almost entirely by refinements in camera and 
printer design ; the second handicap, that of inadequate photographic 
materials, has also become a matter of minor importance. 

The early cameras were rather flimsy contraptions, and the film 
pull-down mechanisms were anything but accurate. Claws of vari- 

* Received October 10, 1936; presented at the Fall, 1936, Meeting at Rochester, 
N. Y. 

** Loucks and Norling Studios, New York, N. Y. 
136 



TRICK AND PROCESS CINEMATOGRAPHY 137 

ous types and spring-operated pawls engaged the sprocket holes. 
Attempts were made to keep the film in registration by the use of 
pressure pads and side shoes, but these were expedients that only 
partly performed that function. It was not until pilot-pin registra- 
tion made its appearance that the problem of accurately positioning 
the film was on its way to a satisfactory solution. 

With the appearance of well-designed cameras, and the introduction 
of dissolving shutters, masking devices, etc., it became easier to pro- 
duce trick effects at the time of shooting, whether in the studio or on 
location. Due to the length of time often required for these opera- 
tions when made on the set, the production of transitional effects, 
such as fades and dissolves, ultimately became a matter for after- 
treatment. Fades were made in the negative by bleaching; dis- 
solves by duping on a printer. Neither method was very convenient, 
nor were the results, especially with dissolves, very satisfactory. 
Intricate work, such as matte shots, still had to be made on an origi- 
nal negative, which was kept in the camera magazine, often for days, 
until all the exposures comprising the composite could be completed. 
Extreme care had to be exercised to overcome variations in exposure 
and contrast. 

With the appearance of suitable duplicating emulsions, it became 
possible to use methods of after-treatment for nearly all the special 
effects required in making the modern motion picture. 

Contrary to the prevailing notion, most trick photography does 
not demand a completely equipped laboratory containing expensive 
and complicated equipment. Most, if not all, the effects used in 
production can be made in the modern camera. The prime re- 
quirement is skill and ingenuity on the part of the operator. A few 
minutes of work with simple tools and odds and ends will often be all 
that is required to make special apparatus for the production of an 
effect. Some time ago a kaleidoscope was needed to produce a mul- 
tiple effect shot of some advertising material. Now such a device 
is quite expensive, and the demand for a kaleidoscope is very limited. 
It took a short time to assemble one out of three pieces of mirror 
held together with film tape. 

Of course, special equipment is essential if speed of production is 
to be attained, and for the more elaborate types of work. Hence, 
in the modern plant will be found optical printers, trick title 
and animation stands, and other devices. The important features 
of these devices will be described later. 



138 



J. A. NORLING 



[J. S. M. P. E. 





Feb., 1937] TRICK AND PROCESS CINEMATOGRAPHY 139 

TRANSITIONAL EFFECTS 

There are many routine steps in process work that can be simpli- 
fied so that operations can be carried on by anyone after brief in- 
struction and practice. Transitional effects, such as fades and 
dissolves, formerly made at the time of shooting the original scene 
on the movie stage, are at present usually made in an optical printer. 
This apparatus affords a wide range in achieving effective results. 
Before describing special apparatus, such as the optical printer, it 
may be of benefit to describe how some of these effects can quickly 
and easily be produced using any motion picture camera. Suppose 
a fade-in (or fade-out) is required. A print (usually made on special 
stock and in accordance with the specifications laid down by the film 
manufacturer for duplicating positives) is rolled up in the magazine 
face-to-face with the required length of negative film (usually a 
special duplicating negative stock, but not necessarily, as any nega- 
tive film can be used provided it is properly developed) (Fig 1). 

It is extremely important to use negative film that has been re- 
wound some days previously so that the emulsion is on the outside 
instead of the inside, as it comes from the manufacturer. In ad- 
dition, it is necessary to keep the positive for a few hours also wound 
emulsion side outward. This procedure imparts a natural curvature 
to the film with the emulsion face on the outside of the curve, and 
assures good contact between the print and the negative in the 
camera gate. (Those who have had experience in making bipack 
color-films will appreciate thoroughly the necessity for preparing 
the films in this way.) A framed leader at the beginning of the 
roll facilitates operations. 

The film is threaded through the camera gate with the print toward 
the lens. Light coming through the lens will expose the negative 
through the print. In threading the positive through the gate 
the print must be framed in relation to the camera aperture. To 
expose the negative, the best light, as well as the most convenient, is 
the reflection from a white card that is evenly illuminated by any 
kind of lamp. Exposure is controlled by setting the lens diaphragm. 

For a fade-in, the camera shutter first is closed, and the camera 
crank turned to the point where the beginning of the fade is to take 
place. Using the single-frame crank facilitates operations, as it 
enables counting by frames. When the desired frame is reached, the 
shutter is opened gradually and at a rate determined by the desired 
length of the fade. 



140 



J. A. NORLING 



[J. S. M. P. E. 





Feb., 1937] TRICK AND PROCESS CINEMATOGRAPHY 141 

For making fades and dissolves in the camera, bipack magazines 
are a great convenience. The best way is to set up the prints in 
separate matched rolls which should have framed leaders spliced on 
at the beginning. Fig. 2 shows two rolls made up; the one in the 
lower compartment of the magazine A, the other B. The marked 
unexposed negative is placed in the upper compartment of the maga- 
zine. After the first exposure through print A has been completed, 
roll B is put into the magazine and the operations are carried out 
as before, except that the shutter must be closed at the point where it 
was opened for the first run. This method of making dissolves is 
well known to most cameramen. The purpose of reviewing it is to 
simplify the description of other, more complicated, methods and 
to make them easily understood. 

SPECIAL EQUIPMENT 

(1 ) The A nimation Stand 

One of the most useful pieces of apparatus in a process or trick 
photography department is an animation or vertical title stand. 
Fig. 3 shows such a stand with a mobile camera carriage which can be 
raised and lowered to accommodate fields of various sizes. This 
particular stand is provided with an automatic focusing device. 
(Fig. 4). 

The long, flat cam is cut to a curve calculated for the focal length 
of the lens. A roller on a push rod follows the curve of the cam, 
which transmits movement to a sliding member, which in turn 
effects movement of the lens. Pins on the lens barrel are engaged 
by slanting slots in the sliding member. 

A mercury vapor M -tube below a window in the table top affords 
a brilliant and even illumination. It is covered with an opal glass 
screen for diffusing the light. The camera, a standard Bell & Howell, 
is driven by a mechanism connected to the single-frame drive shaft. 
This mechanism is designed to allow single-frame exposures to be made 
either continuously or intermittently. The camera speed can be 
varied from 24 to 240 exposures per minute. The average camera 
speed is 160 exposures per minute, but lower speeds are necessary in 
certain kinds of work and when working with extremely slow emul- 
sions such as the 1505 and the newer duplicating negative materials. 

Exposure settings can be made either by varying the lens aperture 
or by changing the shutter opening. For making exposures at a 



142 J. A. NORLING [J. S. M. P. E. 

low speed, or intermittently by one frame at a time, it is extremely 
important to provide some means of maintaining the speed uniform; 
otherwise a variation in exposure between frames may occur. Such 
variations are very slight, but they affect seriously those portions of 
a picture the densities of which are in or close to the underexposure 
region of the H & D curve. A heavy flywheel upon the driving shaft 
of the mechanism assures uniformity of speed. 

The advantage of using the animation camera in making dissolves 
and special dupes lies in the flexibility of control that is possible. 
For split-screen composites, cut-out masks can be used over the 
window. Special screens, lightly shaded in selected areas, can be used 
to hold down exposure in certain portions of the scene. This operation, 
called "dodging," can be used to overcome excessive light values that 
tend to subordinate the central point in a scene. The advantages 
of "dodging" have long been realized in still picture work, but have 
found little application in motion pictures. "Animated dodging" 
can be employed to achieve certain effects ; for instance, it is relatively 
simple to apply this stunt to show a character surrounded by a ghostly 
glow moving about the scene. To produce such an effect in the 
studio would be extremely difficult. The actor would have to be 
wired like an electric sign, and that would probably interfere with 
his work. It would be extremely difficult, if not impossible, to light 
the set properly. 

A great variety of effects can be done on the animation stand, 
among which are the transitional effects known as "wipe-outs." 
Wipe-out is a word that was added to studio parlance several years 
ago to describe the effect obtained when one scene blends into another 
by movement rather than by cross-fading as in a dissolve. Fig. 5 
shows an animation camera set up for making "wipes." Cut-out 
masks consisting of a series of matched pairs are used, numbered for 
convenience. They are laid down over the window one after the 
other, and a single exposure is made in the camera for each mask. 
The masks shown at the left of Fig. 6 contain an opaque section 
moving in from the the left. The mating set, shown at the right, has 
an open space that exactly matches the opaque area in the other. 
One pair of masks is required for each frame of the effect. If the 
wipe-out is two feet long (the most commonly used), thirty-two 
pairs are required, or a total of sixty-four. In making wipes by this 
method, the film is run through the camera twice, the first exposure 
being made through set A, the second through set B. Prints of the 



Feb., 1937] TRICK AND PROCESS CINEMATOGRAPHY 143 

component scenes are loaded in contact with negative stock in bi- 
pack magazines, as described before. 

The method permits many variations of procedure. For instance, 
the effect may start slowly, and be speeded up at the end by eliminat- 
ing some of the masks of the series. An effect can be stopped at 
any point and recontinued. By arranging the masks in different 
ways, either a black space or a lighter area can be used as the dividing 
line between scenes. However, this is a slow and tedious method 
of making wipes, but in special cases its advantages often outweigh 
this handicap. Other methods of making wipes are more widely used. 
A favorite device is a mechanical attachment on an optical printer, 
usually a moving shutter geared to the camera drive so as to move as 
the film runs through. The essential principles are the same as with 
the cut-out masks, but, naturally, the variety of effects is limited. 




FIG. 6. Cut-out masks used for making "wipes" with 
the animation camera. 



A great number of mask patterns is required to meet the increasing 
demands for variety. Fig. 7 shows one page of a catalog listing 
more than 250 different patterns. 

(2) The Optical Printer 

Fig. 8 shows a typical optical printer. Its essential features are: 
(a) a projecting mechanism (which can be a camera rebuilt for the 
purpose); (b) a light-source; (c) a photographing camera; (d) a 
lens designed for unit magnification; and (e) a driving mechanism. 

The projecting camera, in this case, is a Bell & Howell camera, 
rebuilt to allow a beam of light to pass through the aperture. The 
major change required is in the intermittent mechanism. Fig. 9 
shows how the pull-down bar has been changed to an open yoke. 
The pressure pad has been cut away to full-aperture clearance. A 



144 



J. A. NORLING 



[J. S. M. P. E. 



prism is mounted behind the shuttle and looks out through a window 
in the camera door. For convenience in mounting upon the optical 
bench, another right-angled prism is attached to the camera door 

EXPANDING AND CONTRACTING PATTERNS 



E I 



SWING AND SPIN PATTERNS 




FIG. 7. A few mask patterns. 

allowing the lamp house to be mounted to the rear of the camera. The 
lamp house is an Eastman medical spotlight using a 500-watt lamp. 
One of its most important features is the cooling cell with which it is 



Feb., 1937] TRICK AND PROCESS CINEMATOGRAPHY 145 

provided. Heat at the projector gate is a serious problem, as the 
film must remain perfectly flat. In this printer there is no appreciable 
rise of temperature at the aperture. It is necessary that the field 
of illumination be uniformly flat. To achieve this, a thin flashed opal 
glass screen is placed at the front face of the inner prism (Fig. 9). 
The flatness of illumination resulting from this arrangement is entirely 
satisfactory. 

The lens mount (Fig. 10) is built to accommodate lenses of vari- 
ous focal lengths, and is provided with vertical and horizontal ad- 
justments and can be rotated. By shifting its horizontal adjustment 
the lens may be rotated eccentrically. The lens shown in Fig. 10 is 




FIG. 8. A typical optical printer. 

designed for unit magnification. Its equivalent focal length is G 1 /* 
inches and its working aperture is //33 (the aperture is fixed and 
can not be adjusted). Such a small aperture is required for special 
work to be described later. For work requiring more light, lenses of 
larger working aperture may be inserted in the mount. 

The camera and projector drive mechanism is variable-speed, and 
can be run either forward or reverse (Fig. 8). It is provided with 
a heavy flywheel to assure uniformity of exposure. The drive is geared 
to the projector through a splined countershaft and a train of gears 
connected to the single-frame driving shaft in the projector camera 
(Fig. 11). This gear train contains a clutch which can be adjusted 
to drive the projector either forward or reverse, independently of 



146 



J. A. NORLING 



[J. S. M. P. E. 





Feb., 1937] 



TRICK AND PROCESS CINEMATOGRAPHY 



147 



the direction of film travel in the camera. Declutching the projector 
drive allows the projector to be operated by hand. 

The projector camera and the lens are mounted upon separate 
sliding carriages which are adjusted by hand according to the enlarge- 
ment or reduction required (Fig. 8). Calibrated scales indicate the 
size of the picture for each lens and projector setting. 

This optical printer differs from most others in that neither the 
camera nor the projector is adjustable vertically or transversely. 
All horizontal and vertical adjustments of the picture are accomplished 




FIG. 11. Projector camera drive. 

by moving the lens. This method is just as satisfactory in opera- 
tion as the more complicated means required for moving the camera 
or projector. 

Matched and accurate registration of the films in the optical 
printer is of the utmost importance. Lining up for exact registra- 
tion is difficult if one depends entirely upon a scale, no matter how 
fine the mechanical adjustment may be. A very satisfactory and 
convenient method is to line up with target films. The target film 
consists of a pattern of ruled lines and symbols, similar to the pattern 
upon a lens-testing target. A negative of this target chart is placed 
in the projector, and an image thereof projected to a print placed in 



148 



J. A. XORLING 



[J. S. M. P. E. 



an auxiliary aperture plate in the camera. The setting at which 
registration occurs becomes instantly apparent when the image of 
the negative exactly fits the print in the aperture plate. At this 
point the image cancels out the clear spaces in the print, and what 
the operator sees is a gray patch with no trace of the target pattern. 
The production of fades and dissolves by means of the optical 
printer is relatively simple. A print is placed into the projector and 
copied off in the camera and the shutter manipulated to effect fading. 
For making wipes on the optical printer, several methods can be 
used, including the mechanical arrangement previously mentioned. 

This mechanical shutter device is rather 
limited as to pattern variety, for which 
reason film masks are more frequently em- 
ployed (Fig. 12). Mask print A carries one 
side of the matching masks, and mask 
print B the other. These are placed into 
and advanced through the projector ac- 
cording to the data prepared on an ex- 
posure key sheet. The duplicating prints 
that are to be copied are loaded in contact 
with the negative in bipack magazines. 

For sharply defined wipes the lens is 
focused sharply upon the film mask. For 
diffused, or blended-edge, wipes, the focus 
is adjusted to the desired degree of fuzzi- 
ness. This will not affect the picture 
definition hi the slightest degree, since the 
dupe negative is made by contact as de- 
scribed previously, and not by projection. The mask pattern is, of 
course, reproduced by projection. 

A unique attachment for the optical printer is what is called "the 
spinner." Fig. 13 shows a view of this useful mechanism, by means 
of which are produced swinging or turning-field effects. The film- 
advancing mechanism is designed to be rotated upon a vertical 
axis. The device is hand-manipulated, frame by frame, and turned 
through the desired angle as each frame is advanced. The lens, 
having an aperture of f/&, has sufficient depth of focus to keep the 
image reasonably sharp throughout at least 45 degrees of the 90- 
degree turn. The projection head can be adjustable horizontally 
so that the axis of rotation may be centered at any point. Thus it 




FIG. 12. Film masks for 
making "wipes" on the op- 
tical printer. 



Feb., 1937) TRICK AND PROCESS CINEMATOGRAPHY 



149 




Fir,. 13. The "spinner," an attachment for producing swinging or turning-field 

effects. 




1 




FIG. 14. A typical montage. 



150 



J. A. NORLING 



[J. S. M, P. E. 



becomes possible to cause the image to swing in or out at the extreme 
right or left, or to spin about its center. This mechanism is used 
to make "book-page" transitions and other effects that require 
motion to continue as the angle of the scene is changed. 

There are many possibilities for the optical printer in making 
montage and composite-action strips. There are several ways in 
which such make-ups can be arranged. One method will be described 
that was employed in a typical montage containing three different 
scenes combined as shown at D in Fig. 14. Note that the three 




LIGHT 
AREA 



EMULSION 



FIG. 15. Dispersion of light by silver particles in the emul- 
sion, causing change of contrast in optical projection. 

scenes, A, B, and C, are combined in a single frame, D. Note that 
C has been reversed left for right in the composite. 

Masks are made for the areas assigned to the various component 
scenes. The mask for A is inserted into the camera in front of the 
unexposed negative, either as a cut mask placed into the mask slot, 
or as a travelling film mask loaded in contact with the negative. The 
print is put into the projector and the lens moved so that the re- 
quired portion of the original scene shall fall within the open area of 
the mask. Scenes B and C are handled in the same manner, using 
the masks prepared for each. The print of scene C is turned around 
in the projector to reverse the direction of the action as required in 
this particular montage. 

Many special effects are made on the studio stage during the 
filming of the primary scene. For instance, ghosts or visions can 



Feb., 1937] TRICK AND PROCESS CINEMATOGRAPHY 151 

be made to appear in the scene by photographing the primary scene 
through a partially silvered mirror. The ghost image is picked up 
by reflection in the mirror from a set placed at right angles to the 
optical axis of the camera lens. The same effect can be achieved by 
after-treatment in the optical printer, and with greater ease. 

Glass shots are seldom used at the present time although they were 
quite the vogue several years ago. The background and sometimes 
part of the foreground are painted upon glass in such a manner as to 
match adjoining parts of the scene. Better results can be attained 
in a simpler manner with the optical printer. 

Matte shots often give the cameraman considerable trouble. Such 
work can easily be done with the optical printer. In fact, the optical 
printer is among the most versatile tools at the command of the 
film technician. 

Print quality is an important matter in many types of trick photog- 
raphy. It is unnecessary to point out that prints must be clean and 
free from blemishes of any kind; but the degree of contrast and the 
density of the duplicating print must receive special consideration, 
depending upon whether the prints are to be copied by projection or 
by contact. If dupes are to be made by projection, the prints should 
be much less contrasty than for contact duping. The reasons for 
this are well known, but it will do not harm to give a short review of 
them. 

A change in contrast takes place in optical projection. Fig. 15 
shows diagrammatically what ocurs to light passing through the 
print emulsion. The light is dispersed by the silver particles in the 
emulsion. The dispersion is greater where the deposit of silver is 
heaviest, as in the shadows of a picture, and less in the clearer por- 
tions of the picture, as in the highlights and bright objects. 

In contact duping this dispersion has little effect upon the con- 
trast; but in projection, this scattering of light will result in a con- 
siderable increase of contrast. The aperture of the lens gathers very 
little of the dispersed light, even when using large-aperture lenses of 
short focal length. When stopping down, still more contrast is ob- 
tained, because less of the dispersed light is gathered by the aperture. 
In optical printing, if the density of the shadows in the print is great 
in relation to the lighter portions, too much dispersion takes place to 
retain much shadow detail. Little dispersion occurring in the lighter 
portions will result in full exposure at those points as against under- 
exposure in the shadows. Any change of aperture setting will affect 



152 J. A. NORLING [J. S. M. P. E. 

the contrast to some degree, and if the range in setting is large, as 
from //4 to //32, a marked increase of contrast will occur. With 
this consideration in mind, exposure adjustments should be made by 
varying the intensity of the light or the duration of exposure, and 
not by changing the lens aperture, as is done in many optical set-ups. 

There is no better way to insert a background in composite photog- 
raphy than by background projection or by the Dunning process, 
but some composite scenes requiring background insertion can be 
made on the optical printer at less expense, if the picture action is 
suitable: for instance, a mass of clouds, or the sun rising over a 
distant mountain, or any silhouette background across which there 
is no action. Some background insertions can be made without 
preparing any travelling masks, but in most cases travelling masks 
are found necessary 

While comparatively little footage is now being made with travel- 
ling masks, the use of travelling mattes affords an opportunity for 
achieving some striking results and is peculiarly adaptable to certain 
kinds of composite work. Travelling masks are often difficult to 
make, especially when they have to be animated or drawn frame by 
frame to fit the action or other elements of a scene 

Among the interesting possibilities of travelling masks is the pro- 
duction of combined cartoon and actual photography. To produce 
such animated composites it is necessary to have some means of 
accurately confining the cartoon figure action in relation to actual 
objects in the scene. Each frame of the primary or background 
scene is traced off to furnish a guide for the drawing of the cartoon 
creature. 

There may be several ways of doing this, but the most convenient 
method is to use a projection camera in the animation stand (Fig. 16). 
Each frame of the original scene is projected down upon a sheet of 
paper accurately registered on standard animation pins. The artist 
can easily trace the outline of the figure upon the projected image. 

The tracings are used as guides for the animator, who draws the 
cartoon character directly upon the same sheet. The pencil drawings 
are then traced in India ink upon transparent sheets of celluloid, 
which are opaqued on their reverse sides in various shades of opaque 
paints. Travelling masks are prepared by photographing these 
drawings and multiple-exposing over the primary scene by several 
runs through the optical printer. 

Fig. 17 shows an animation stand set up to produce cartoons by 



Feb., 1937] 



TRICK AND PROCESS CINEMATOGRAPHY 



15:* 




FIG. 16. Projection camera used with animation stand, 
for producing combined cartoon and actual photography. 




FIG. 17. Animation stand set up for producing cartoons 
by the Technicolor process. 



154 



J. A. NORLING 



[J. S. M. P. E 



the Technicolor three-color process. The wheel located directly 
below the lens contains the color-separation niters: red, green, and 
blue. Exposures are made of each drawing successively through the 
three filters, which are balanced for the negative emulsion used anc 
for the characteristics of the lamps employed. This provides 
single negative instead of three separate negatives, but three time: 
as long as a negative for black-and-white. Three matrices, one foi 




FIG. 18. 



A trick scene containing an animated drawing combined with 
actual scenes. 



each color, are made from this negative on an optical "skip printer. 
Prints are made from the matrices by the Technicolor imbibitio 
process. Wipes, dissolves, and multiple exposures for color work ai 
easily produced on either the cartoon camera or the optical print( 
from duplicating prints as described for black-and-white worl 
Both the animation camera and the optical printer are provided wit 
switching arrangements that can be pre-set so that three exposure 
are made every time the exposing button is pressed. This mechanisi 
is designed so that, no matter what the operator does or fails to di 



Feb., 1937] TRICK AND PROCESS CINEMATOGRAPHY 155 

it is impossible to get out of synchronism by exposing more than three 
or less than three frames for each impulse. 

The animation stand is particularly advantageous in photographing 
special titles. Some of the effects seen in the elaborate titles in 
modern films require considerable work of many different kinds. 
Special lighting, such as by spots, etc., is often employed. A very 
Effective stunt used for the main title of an industrial film required 
'that a shallow tank of water be used above the title cards and that 
the transitions between the titles be effected by a disturbance on the 
Surface of the water. Clay models, small sets, etc., often required 
as backgrounds, can be photographed on this type of animation stand. 
Panoramic boards that can be adjusted to move in varying steps, 
cither horizontally or vertically, are provided for special trick title 
manipulation. 

Probably the most difficult, and certainly the most tedious and 
painstaking to make, of all trick scenes, are those containing animated 
drawings combined with one or more actual scenes. An illustration 
of the steps involved in a typical composite of this kind is shown in 
Fig. 18. It was desired to combine the following elements: A, a 
view of a telephone switchboard; B, a view of a revolving globe; 
C, animated drawings of transoceanic and transcontinental radio 
telephone circuits. The final composite is shown in D. 

The scene starts with a full view of the switchboard. The camera 
then tilts upward and swings to place the switchboard down in the 
lower right-hand corner of the frame. As this swing takes place, 
the revolving globe containing the animated circuits is disclosed as 
if it were behind the board. The following steps were required in 
producing this composite: 

(1) Scene A, that of the switchboard, was photographed, and a duplicating 
print made thereof. 

(2) This print was projected down upon the field in the animation stand, and 
using this projected image as a guide, a series of masks was made to cover up the 
portion of the frame to be occupied by the globe. Since the footage consumed by 
the downward movement of the switchboard amounted to fifteen feet, this meant 
that 240 matched pairs had to be made. 

- (5) The globe was mounted upon a turntable and photographed as it rotated. 
It was necessary to provide guide lines upon the globe to be followed by the ani- 
mator when drawing the animated radio circuits. These lines were not to show in 
the finished composite, and this presented what seemed at first quite a problem. 
It was solved by drawing the lines upon the globe in red, and loading a bipack 
camera with clear-base panchromatic negative film in front and orthochromatic 
film in back. Lighting and photographic tests were made until a balance was 



156 J. A. NORLING [J. S. M. P. E. 

reached such that the front negative showed no trace of the lines. The lines were, 
of course, reproduced as nearly clear lines upon the back negative. A print from 
each of these negatives was made in the optical printer. 

(4) The print from the ortho negative of the globe was placed in the animation 
stand projector and the guide lines traced off upon white paper sheets registered 
upon pins in the animation field. 

(5) The pencil tracings were then used as guides for the animation, which was 
done in India ink upon sheets of celluloid. Notice that the animation lines upon 
the composite are white, but that the drawings are black. 

(6) These black line-drawings were placed over a white background, and a 
negative made thereof upon positive stock exposed for extreme contrast and 
developed to a high contrast. 

(7) This negative was loaded in the animation camera in contact with duplicat- 
ing negative stock, which was exposed by the light from the window in the field 
as one set of the masks obtained in a previous operation was laid down over the 
light-source. This gave the first exposure on the final composite negative. 

(8) The next operation required a similar treatment, using the same masks with 
the print of the globe (this is where the print obtained from the panchromatic 
negative of the globe was used). This operation gave the second exposure on 
the composite negative. 

(9) In the next operation, the duplicating print of the switchboard was set up 
in contact with the negative, and exposures were made as for operations 7 and 8 
except that the mates of the masks were used. 

From the above-given very abbreviated resume of a typical com- 
plex scene it can be appreciated that accurate registration plays a 
leading role in process photography. Accurate registration re- 
quires placing and holding the film exactly during exposure, but it 
is equally important that there be no movement of the cameras 
when taking any of the component scenes. All takes must be made 
with the cameras rigidly mounted. Hand-driven cameras are likely 
to move slightly at each turn of the crank, and even the slight move- 
ments produced through this cause result in unsatisfactory registra- 
tion. 

It is always advisable to have some means of checking the accuracy 
of registration. For this purpose a very simple procedure is used. 
A target chart is prepared containing ruled lines. The camera is 
placed in the animation stand and the chart is photographed in the 
following manner : First an exposure is made of the right half of the 
chart, covering up the left half with a black screen. Then the 
film is turned back to the starting point, and a second exposure is 
made while covering the right half and exposing the left half. Ex- 
posures are made in this way of the top and bottom parts. Exposures 
are made also by running forward while one-half is exposed, and 



Feb., 1937] TRICK AND PROCESS CINEMATOGRAPHY 157 

making the second exposure while running the camera in reverse. 
The lines join up perfectly, and when projected, exhibit no move- 
ment apart, if the mechanism is in good working order. If there is 
any movement visible upon the screen at the juncture of the lines 
it is an indication that the pilot pins have become worn and should 
je replaced. 

Film shrinkage is a real handicap, and all prints made for composite 
work should have exactly the same dimensions in the perforations. 
[f old and shrunken films have to be used in special cases, provision 
must be made accordingly. All prints should be made on machines 
that will not introduce new errors in registration. It is sometimes 
advantageous to make one's own prints in the camera to assure ac- 
curacy of registration. This is unnecessary where first-class labora- 
tory service is available. 

Process photography long has been and continues to be a valuable 
element of film production. Several volumes would be needed to 
review the subject completely. This paper has been limited to 
certain basic processes and brief descriptions of special equipment and 
typical production procedure. The processes described will, it is 
loped, be of some value in suggesting other methods and improve- 
ments in technic. 



SLIDE-RULE SKETCHES OF HOLLYWOOD* 

H. G. TASKER** 



Summary. A brief review of some of the current improvements in production 
methods and production tools used in the motion picture industry. 

The subjects mentioned include processing laboratories, cameras, synchronizing 
methods, and various items of sound equipment and methods, including a review of the 
present status of push-pull recording. 



The paragraphs that follow comprise brief and informal glimpses 
of some of the technical changes that have come over Hollywood 
and its production methods principally within the last year. The 
broad subject of technical advances in Hollywood production methods 
is by no means fully covered, the individual "sketches" are by no 
means complete, and even their accuracy may be doubted; but it is 
hoped that they will afford to many motion picture engineers re- 
siding in other parts of the world a measure of familiarity with what 
is going on in Hollywood. 

Double-Track Recording. Push-pull recording has at last become an 
actuality in a number of Hollywood studios. The Society is already 
familiar with the remarkable results obtained by RCA Manufacturing 
Company with its push-pull recording of the Class B type first demon- 
strated at the Atlantic City Convention of the Society (April, 1934), 
and since shown in improved forms at several later Conventions and 
described in the JOURNAL. Perhaps we are less familiar with the 
push-pull system of Electrical Research Products, Inc., which was 
developed out of early work of Douglas Shearer (Metro -Goldwyn- 
Mayer) and was shown at the Hollywood Convention in May, 1935. 
It comprises two variable-density tracks, of the Class A type but 
arranged in phase opposition. The method does not have the in- 
herent noiselessness of the RCA method, but owing to a secondary 
effect it nonetheless accomplishes a very substantial noise reduction. 

* Received December 1, 1936; presented at the Fall, 1936, Meeting at Roch- 
ester, N. Y. 

'* Universal Pictures Corp., Universal City, Calif. 
158 



SLIDE-RULE SKETCHES OF HOLLYWOOD 159 

This secondary effect is the fact that when noise reduction is applied 
to a push-pull track, noise reduction "thumps," being in phase on 
the tracks, are cancelled out by the push-pull connection of the 
system. Consequently, higher-speed, lower-margin noise reduction 
may be successfully applied to this push-pull system than was possible 
in the former single track. Furthermore, the push-pull arrangement 
cancels out much of the second harmonic arising from photographic 
distortions, making possible the use of much greater noise reduction 
on the one hand, and considerably higher maximum modulation on 
the other. As a result of these considerations the ERPI push-pull 
method obtains about the same extension in volume range as does 
the RCA method. 

First releases of sound originally made by the variable-density 
push-pull system and dubbed to single track were made by Metro- 
Goldwyn-Mayer and include The Great Ziegfeld. First release of 
sound-track made with the ERPI push-pull system is the musical 
score for Magnificent Brute by Universal. The latter company has 
also the first ERPI all push-pull picture in production: Top of the 
Town, a musical, to be released in January. No releases have yet 
occurred containing sound originally made by the RCA push-pull 
system. 

Novel and Economical Editing Theaters. The average major studio 
.n Hollywood has from fourteen to twenty projection rooms, or review 
rooms. The bulk of these are used by the film editors, who do their 
detail work with Moviolas but must have something more like the 
theater to get the "feel" of complete sequences as they are made 
ready. There must also be such places for the chief editors, directors, 
and others to see the progress of the cutting. Universal has devised 
an exceptionally economical and effective editorial projection room by 
installing projection type Moviolas with filtered incandescent lamps 
land blower gates. This makes it possible for the film editor to be his 
pwn operator, saving several hundred dollars in salaries each week. 
JHe first threads the machine, which is located in a fire-proof booth, 
iand then walks into the review room proper, from which the machine 
may be operated by remote control . The fact that these Moviola pro- 
Sectors can be started, stopped, or reversed at the touch of a button 
i>aves a tremendous amount of time in studying difficult cuts. 

Synchronizing Methods. Studios employing interlock motor drives 
'or camera and sound recorder have found that the old method of ap- 
3lying synchronizing marks to the two films before rotating the ma- 



160 H. G. TASKER [J. S. M. p. E. 

chinery loses much valuable time. Several studios have abandoned 
these synchronizing marks, and use, instead, a method that consists 
in clapping a pair of sticks together in front of the camera after the 
motors are up to speed. These sticks are usually combined with the 
camera slate. The result is by no means wholly satisfactory, as time 
is still lost while the assistant cameraman gets out of the scene with his 
slate. Furthermore, the sound of the clap-sticks is often annoying to 
the actors, making it difficult for them to get into the mood of the scene. 
Severe trouble of this sort is encountered when working with animals, 
which are often frightened by the sound of the sticks. As a result 
of these difficulties, and through the efforts of the Sound Committee 
of the Academy of Motion Picture Arts and Sciences, a standard 
synchronizing method is evolving out of early work of Metro- 
Goldwyn-Mayer and others. In this system a portion of the picture 
negative is fogged by a neon light, and at the same instant the 
exposure of a portion of the sound-track is varied so as to print upon 
each film synchronizing marks anywhere from 6 to 12 inches long. 
This act will be unaccompanied by any sound other than a buzzer 
of moderate tone, to indicate to the director and cast that the equip- 
ment is up to speed and that action may begin. Metro -Goldwyn- 
Mayer and United Artists have made provision also for slating 
between scenes by providing a cameraman with a button for starting 
the cameras instantaneously at any time without communicating 
with the sound recording room. Other companies are following 
suit. While all this may seem comparatively unimportant, it must 
be remembered that the slating and synchronizing operation occurs 
at the moment when everything is in readiness for a take, and 
consequently any delay at this moment delays the entire company 
and is very costly. It is estimated that if used by all studios, the 
new slating and synchronizing methods will save approximately 
$250,000 per year. 

New Microphones. Two new microphones have made their ap- 
pearance in production: a small non-directional, or "cue ball" micro- 
phone, by ERPI, and the unidirectional ribbon microphone by RCA. 
The former is used for all current productions by General Service 
Studios, Paramount, and Universal. Those who use it feel that its 
characteristics are smoother than those of the former dynamic micro- 
phones, but all apply some corrections to its characteristics. Its 
non-directional characteristic is, in part, a benefit and a detriment. 
Ideally, a microphone should be absolutely non-directional over the 



Feb., 1937] SLIDE-RULE SKETCHES OF HOLLYWOOD 161 

angle from which wanted sound may be received and absolutely 
non-responsive from all other directions. No such microphone is 
available, although the unidirectional ribbon of RCA is an approach 
to the solution of this problem. Columbia has employed it in 
recording Pennies from Heaven, the new Bing Crosby picture. If 
reduced in size and increased in sensitivity it would lend itself to 
still greater usefulness. 

Millions in New Sound Equipment. By far the most exciting 
change that is taking place in Hollywood just now is a complete 
remodelling and modernizing of the sound equipment in every 
studio. This major upheaval is caused in part by the advent of 
push-pull recording and of improved sound apparatus, all the 
way from the microphone in the studio to the loud speaker in the 
theater; and also in part by the tremendous increase in studio ac- 
tivity and the very hopeful outlook for the future. A still greater 
contributing factor is a new and very liberal merchandizing policy 
on the part of the two major licensing companies in the sound re- 
cording field, each of whom has offered to its licensees an arrangement 
whereby new equipment and the modernization of old equipment 
may be paid for out of royalties. The result is a tremendous program 
of modernization and expansion in sound apparatus averaging more 
than $200,000 per studio. 

New Cameras. The bulk of Hollywood productions have, over a 
period of years, been photographed with Mitchell cameras, and the 
manufacturer of this product has introduced a comparatively quiet 
camera known as N C, which has enjoyed considerable favor during 
the past two years. More recently Twentieth Century-Fox Film 
Corp. designed their own silent camera, used for all productions of 
that company but not elsewhere. Its interesting features were 
described in the Report of the Progress Committee of the Society, 
in the July, 1936, JOURNAL. Within the past two months the same 
company has sponsored the first appearance in Hollywood productions 
of the DeBrie camera, manufactured in Paris, which has a number of 
very unusual features and promises to become quite popular as a 
result. In contrast with former cameras, which were enclosed 
within blimps so as to reduce camera noise, or in which the mechanism 
was built to operate as quietly as possible in the hope of avoiding 
the use of the blimp, the DeBrie camera is so designed that its outer 
case is, in fact, a blimp, within which the mechanism is relatively 
exposed. As a result of this design, the camera produces less noise 



162 H. G. TASKER [J. S. M. P. E. 

than a regular Mitchell camera in a blimp, and still less noise than 
the N C Mitchell camera without a blimp. At the same time its 
overall size and weight are less than that of the average camera in 
a blimp and its mobility and controllability are considerably ex- < 
tended. A feature of great interest to cameramen is the ability to 
view the action continuously through the film itself during a take. 

Head-Turning Equalizer. Since the microphone is inherently a 
device that should be placed close to the source of sound, the present 
technic is to support it on a microphone boom provided with many 
mechanical gadgets to enable rapid shifts of position of the micro- 
phone in and out, up and down, left and right, and rotational. 
The microphone boy must be a wizard of dexterity, equipped with 
a sixth sense of anticipating unrehearsed movements of the actors. 
Despite his skill there are often situations with which he can not 
cope; as, for example, when an actor's head is turned from the 
camera for so short an instant that the microphone boy is unable to 
follow him, or when the actor stoops down so low that he can not be 
followed for fear that the microphone will appear in the picture. 
The result is a loss of high frequencies and, consequently, of in- 
telligibility. To cope with these situations Universal has developed 
what they call a "head-turning" equalizer, which at a simple twist of 
the mixer's wrist restores the high frequencies to their normal balance 
and makes it possible for him to follow instantaneous movements of 
the actors. 

New Processing Laboratories. There has been considerable ac- 
tivity in constructing new laboratories and modernizing old ones. 
Consolidated Film Industries has just opened a fine new plant, effi- 
ciently arranged and thoroughly equipped. Solution tanks are un- 
usually large and there is adequate provision for agitation of the solu- 
tion during development. The plant is so arranged that raw stock 
enters one end of the building, passes successively through the print- 
ing, processing, and projection rooms, and is delivered to the ship- 
ping room at the other end of the building. This laboratory will 
include a complement of non-slip sound printers. 

Warner Bros. Pictures, Inc., are building a new processing labora- 
tory on the Burbank lot. Abandonment of the Sunset Laboratory 
is contemplated. 

International Cinema, Ltd., has opened a fine new plant, of which 
an important feature is the photoelectric control of various steps of 
printing and processing,;: 



Feb., 1937] SLIDE-RULE SKETCHES OF HOLLYWOOD 163 

Pre-Selection vs. the Split- Film Method. At least one studio that 
had been using 17 VV mm. film for sound recording has found it more 
economical to return to 35-mm. film and employ, instead, the pre- 
selection method, with the result that less processing is done but 
much more film is used. The ITVVmm. method consisted in re- 
cording along one edge of the 35-mm. film, then reversing the film in 
the magazines, and recording along the opposite edge. The film 
was processed in the 35-mm width, and then split before printing. 
The sound dailies were also on 17Vrnim. an( j the resulting savings 
for an average large studio amounted to around $4000 a month. 
In practicing this method all sound negative must be processed. 

In the pre-selection method the director designates OK takes 
and Hold takes, all others being marked NG. All but the OK 
takes are broken out of the roll before processing. Since only one 
edge of the film has been recorded upon, the NG takes (and later the 
Hold takes) may be spliced together and the opposite edge used for 
printing sound dailies; thus effecting two economies, in that the NG 
negative is not processed and the daily print stock is obtained without 
cost. These economies alone amount to around $3500 a month 
for an average studio, and nearly offset the economies obtained by 
the ITVj-mm. method. There remains, however, a reserve of 1 to 
2 millions of feet of NG stock per year per studio. New ways of 
using this stock, including leaders, effects negatives, effects positives, 
and even action dailies, have made the pre-selection method defi- 
nitely more economical than the split-film method 



EFFECT OF LIGHT-SOURCE SIZE WITH 16-MM. 
OPTICAL SYSTEMS* 



G. MILI** 



Summary. An investigation of two spherical and aspheric condensers, with 
an f/ 1.65 projection lens and a test series of high-intensity biplane filament lamps 
ranging from 200 to 2000 watts, has been conducted to determine the levels of illumina- 
tion attainable in 16-mm. projection by varying the light-source size. 

An analysis of spherical and aspheric condensers of various designs 
to show the best possible combination with an //1. 65 lens for 16-mm. 
projection has already been presented. 1 This study was made with 
the biplane filament light-source now most widely used in the field, 
namely, a 500-watt, 100-volt lamp, the light-source dimensions of 
which, when backed with a spherical mirror, are about 9 by 9 milli- 
meters. Since the light-source requirements vary with the spacing 
between the condenser and the film-gate, it was considered advisable 
to continue the study, employing light-sources of various dimensions. 

PROCEDURE 

A series of tests was run with spherical and aspheric condensers 
65 mm. in diameter and having a magnification of 2.08. A sketch of 
the optical elements involved is shown in Fig. 1, while the dimensions 



TABLE I 


Condenser Data 


Diameter 

(mm.) 


E.F. 
(mm.) 


Magnification 
X 


A 

(mm.) 


B 

(mm.) 


65 


33.2 


2.08 


34.5 


87.5 


65 


33.4 


2.08 


32.3 


85.7 



Curvature 

Aspheric 
Spherical 

for various spacings are given in Table I. The dimensions and the 
electrical ratings of the filaments, which, when used in conjunction 

* Received October 9, 1936; presented at the Fall, 1936, Meeting at Roch- 
ester, N. Y. 

** Westinghouse Lamp Division, Westinghouse Elec. & Mfg. Co., Bloomfield, 
N.J. 
164 



LIGHT-SOURCE SIZE WITH 16-MM. SYSTEMS 



165 



with a spherical mirror, formed the four light-sources used for test, 
are given in Table II. Fig. 2 shows the appearance of the sources 
when lighted. 



Watts 

200 

500 
1000 
2000 

*Approximate. 



TABLE II 
Light-Source Data 

Filament 
Volts Construction 

100 CC-13D 

100 C-13D 

100 C-13D 

100 C-13D 



Source 
^Dimensions 

(mm.) 

5 X 5 

9X9 

12 X 12 

15 X 15 



The total light output and the brightness at the center and the four 
corners of the screen were measured, varying the position of the 
film-gate from the theoretical plane of the light-source image to the 



LIGHT SOURCE 





CONDENSER 



LIGHT SOURCE IMAGE 




FILM GATE I \ 



REAR ELEMENT 

FRONT ELEMENT 
PROJECTION LENS 



FIG. 1. Optical elements of 16-mm. film projection system. 

point where the gate was located against the condenser surface. This 
procedure was followed with both condensers. The details of the 
method of taking the readings and photographs of the instruments 
involved have already been published. 1 

Through back-testing, the effective light-source size to fill com- 







200-W 500-W 1000-W 

FIG. 2. The test light-sources. 



2000-W 



166 



G. MILI 



[J. S. M. P. E. 



\ 




1.0 .8 . .4 20 

RELATIVE DISTANCE FROM LIGHT-SOURCE IMAGE PLANE TO FILM-GATE 

FIG. 3. Screen light output data for various light-sources with spherical 
condensers (65 mm. in diameter, 2.08 X magnification). 



fi 2.C 




6 ,6 A .2 

RELATIVE DISTANCE FROM LIGHT-SOURCE IMAGE PLANE TO FILM-GATE 

FIG. 4. Screen light output data for various light-sources with aspheric 
condensers (65 mm. in diameter, 2.08X magnification). 



Feb., 1937] LIGHT-SOURCE SlZE WITH 16-MM. SYSTEMS 



167 



pletely the exit pupil of the system, for various spacings between the 
film-gate and the condensers, was determined. 

RESULTS 

The curves in Figs. 3 and 4 show the variation in screen light out- 
put for the spherical and aspheric condensers, respectively, with 
the gate position varied from the plane of the filament image to the 
condenser surface. The full-line portion of the curves represents the 
region in which the screen appearance is free from light-source images 
and striations. It may readily be seen that the increase in the light on 
the screen is less than the increase in lamp wattage when the gate 
position is near the plane of the light-source image. On the other 



I 







V 




























\ 


\ 




























\ s 


\ 


, 


A, 

- s 


PHEB 

=HtR 


1C C 

CAL 


ONDt 

CONC 


NitW, 

ENSE 


s 














\ 


\ 




























*v 
N, 


\ 




























\ 


\ 










* 


















\ 


\ 








i 




















\ 


\ 






I 






z 

a 
















^\ > 
\ 


\ 




t 




















*S \ 

X 


X 


\ 






i 

1 






















o 

i 






G 






















\ 

































RELATIVE DISTANCE FROM LIGHT-SOURCE IMAGE PLANE TO FILM-GATE 

FIG. 5. Variation in light-source requirements 
for various positions of the film-gate. 

hand, the increase becomes more nearly proportional to the wattage 
the closer the film-gate is brought to the condenser. This is ex- 
plained by the fact that the light-source requirements of the optical 
system differ with the position of the film-gate. 

As may be noted from Fig. 5, the maximum dimension of the ef- 
fective light-source is larger, the closer the film-gate is placed to the 
condenser. Now, with the film-gate near the plane of the filament 
image, the maximum dimension of the effective light-source is about 
6 millimeters; and it is evident from Table II that, with the excep- 
tion of the 200-watt lamp, all the test lamps more than fill the exit 
pupil of the system. The increase in screen light output is therefore 
caused by the fact that with the larger sources, only the central por- 



168 



G. MILI 



[J. S. M. P. E. 



tion of the filament is effective, and this is brighter than the outer 
regions. This may be seen in Fig. 6, which shows an image of the 
type of light-source used for test. Upon this image are superimposed 




HORIZONTAL BRIGHTNESS GRADIENT (H-H,j 



- VERTICAL BRIGHTNESS GRADIENT, BACK COIL (V-V,) 
....... VERTICAL BRIGHTNESS GRADIENT, FRONT COIL(v 2 -Vj) 



FIG. 6. Curves of the periodic variation, or brightness 
gradient, across and throughout the length of a front and rear 
coil of a biplane filament with a spherical mirror, are super- 
imposed upon an image of the light-source. A scale drawing 
of the light-sensitive cell used for the measurements correlates 
the size of the test spot with the coil diameter. 

the curves of the brightness gradient across the filament and through- 
out the length of a front and rear coil. 

As the film-gate is moved closer to the condenser, and, therefore, 
farther from the plane of the light-source image, the size of the ef- 
fective light-source becomes greater than sufficient to accommodate 
the sources involved. At and beyond this point, except for the secon- 



Feb., 1937] LIGHT-SOURCE SlZE WITH 16-MM. SYSTEMS 



169 



dary effect of aberration and slight test errors, the screen light out- 
put should increase in the same ratio as the lamp wattage. 

In order to bring out the effect of source size and brightness gradi- 
ent, the screen light output curves are plotted on the basis of equal 
efficiency for all lamps. In practice, however, a comparison on the 
basis of equal-life performance would be more important. For a given 
filament construction and equal life, lamp efficiency increases with 
wattage because of the increase in wire size. This is indicated in the 
curve plotted in Fig. 7. Accordingly, a somewhat greater increase 



132 



UJ 



r 

UJ 

cr 



400 



600 1200 

LAMP WATTAGE 



1600 



2000 



FIG. 7. 



Increase in efficiency for equal life with increase in wattage 
(100-volt biplane filament projection lamps). 



in screen light output than shown in Figs. 4 and 5 would be afforded 
for equal life if the higher-wattage lamps were employed. 

Curves of the corner-to-center screen brightness ratio are plotted 
for the spherical and aspheric condensers in Figs. 8 and 9, respectively. 
It is evident that with spherical condensers, because of the effect of 
spherical aberration, satisfactory uniformity is achieved only with 
the larger light-sources. For small sources a dark center occurs at 
the middle positions of the film-gate, thereby making the corner- to- 
center screen brightness ratio greater than unity, as is indicated in 
the curves obtained with the 200- and 500- watt lamps. Except for 
positions of the film-gate near the plane of the light-source image, 
the screen brightness curves with the aspheric condensers are very 
uniform and spaced very closely together. This discloses the fact 
that, with aspheric condensers, on account of less residual aberration, 



170 



G. MILI 



[J. S. M. P. E. 



- 2 XX) 



>i- 



M 



RELATIVE DISTANCE FROM LIGHT-SOURCE IMAGE PLANE TO FILM-GATE 

FIG. 8. Screen uniformity data for various light-sources with spherical con- 
densers (65 mm. in diameter, 2.08X magnification). 



.4 .a 

RELATIVE DISTANCE FROM LIGHT-SOURCE IMAGE PLANE TO FILM-GATE 

FIG. 9. Screen uniformity data for various light-sources with aspheric 
condensers (65 mm. in diameter, 2.0SX magnification). 



Feb., 1937] LIGHT-SOURCE SlZE WITH 16-MM. SYSTEMS 



171 



the screen brightness uniformity is largely independent of the light- 
source size. 

For the purpose of supplying a more direct comparison, values of 
the screen light output and the brightness ratio, with the film-gate 
at the point where screen striations disappear, are listed in Table III. 
Some of the figures presented for the 500-watt lamp in this table are 
slightly different from figures already published, 1 check tests having 
warranted these revisions. 



Condenser Data 
Curvature Distance* 
(B- C)** 
(mm.) 

Spherical 45 . 7 



TABLE III 
Projection Data 

Effective Light-Source 
Dimen- Shape 

sions 
(mm.) 



13.7 



Round 



Aspheric 25.0 10.8X11.0 Oval 



Screen Illumination Data 


Light- 
Source, 
Watts 


Relative Brightness** 
Light Ratio 
Output Corner-to-Center 


200 


0.44 


1.37 


500 


1.02 


1.11 


1000 


1.70 


0.84 


2000 


2.60 


0.81 


200 


0.63 


0.75 


500 


1.30 


0.77 


1000 


1.90 


0.76 


2000 


2.82 


0.76 



CONCLUSIONS 

It appears from a survey of the test results, Table III in particular, 
that large increases in screen illumination may be attained with both 
spherical and aspheric condensers by increasing the wattage of the 
lamps, even when the light-source size exceeds that required to fill 
the exit pupil of the optical system. A higher screen light output with 
satisfactory screen uniformity may be attained with the aspheric 
than with the spherical condensers, with all the test light-sources. 
However, this advantage is less marked with increasing light-source 
size. The foregoing conclusions apply not only to the two systems 
under test, but also in a proportionate measure to the optical sys- 
tems previously tested, 1 making use of 26-mm. and 45-mm. con- 
densers. 



* Represents separation between film-gate and plane of light-source image 
necessary to eliminate screen striations (see Fig. 1). 
** At equal lamp efficiency. 



172 G. MILI 



REFERENCE 

1 MILI, G., AND COOK, A. A.: "Condensers for 16-Mm. Optical Systems," /. 
Soc. Mot. Pict. Eng., XXVI (June, 1936), No. 6, p. 603. 

DISCUSSION 

MR. CARLSON: How do the levels of illumination obtained with this experi- 
mental system compare with the better systems in commercial use today? I 
notice, for example, that the space between the light-source and the condenser 
seems to be appreciably greater in the system used for test than is common in 
present practice. Is, then, the efficiency of light utilization of this experimental 
system higher or lower than that of present systems? 

MR. MILI: The efficiency of utilization, or, to be more specific, the screen 
lumens per watt, is higher with optical systems employing closely spaced conden- 
sers and light-sources. However, the short distance between the light-source and 
the condenser limits the wattage of the lamp that may be employed, and this in 
turn limits the level of illumination attainable upon the screen. On the other hand, 
even though the efficiency of the larger systems is less than that of the smaller 
systems, it is possible to use lamps of higher wattages and thereby attain higher 
levels of illumination. 

MR. CARLSON: I should like to have more specific information on the level of 
illumination that could be achieved with a 2000-watt light-source and the system 
described here, as compared with the most efficient commercial optical system 
employing a 750-watt light-source. 

MR. MILI: A highly efficient optical system available commercially for the 
750-watt lamp makes use of a 45-mm. aspheric condenser system. This system 
would be about 10 per cent more efficient than the 65-mm. condenser employing 
the same light-source. Now, let us assume that the 65-mm. condenser system could 
accommodate a 2000-watt lamp, whereas a 750-watt lamp is the maximum light- 
source that might be used with the 45-mm. condenser. On the basis of equal 
lamp life, the relative screen lumens attainable with the two systems would be 
approximately 3.15 and 1.90. 

MR. JOY: How could the efficiency of the system be improved to obtain even 
higher screen illumination levels? 

MR. MILI: The tests of the optical systems were made with an //1. 65 pro- 
jection lens, which is at present the fastest lens available. Higher illumination 
values could be attained either by increasing the aperture of the projection lens 
or by increasing the filament brightness and reducing the life of the lamp. 

MR. CRABTREE: Nothing has been said about the cost of aspheric condensers 
as compared with spherical. 

MR. MILI: Aspheric condensers are more expensive than spherical, although 
I do not know in what proportion. I notice, however, that there is an increasing 
tendency toward the use of aspheric condensers because of their better perform- 
ance, which leads me to believe that the cost of the condensers, either spherical or 
aspheric, is small compared with the cost of the complete projector. 

MR. TASKER : The cost of aspheric condensers is roughly twice that of spheri- 
cal, but, as has been said, it is a relatively small part of the total cost of the 
projector. 






MEDICAL MOTION PICTURES IN COLOR' 
R. P. SCHWARTZ** AND H. B. TUTTLEt 



Summary. Improvements made during the past year in methods, apparatus, and 
materials used in making medical motion pictures, particularly Kodachrome, and 
the characteristics of an emulsion suitable for exposure with artificial light are dis- 
cussed. The uses of special accessories for medical motion picture photography are 
described. 



Increasing use of motion pictures as a part of clinical and surgical 
records has followed each successive improvement intended to meet 
recognized requirements. Only in the past few years have the re- 
maining essentials for the field of medicine and surgery been fully 
met. 

Since the first 16-mm. medical motion pictures were made in 1922, 
a continuous effort has been made to attain this objective. A rec- 
ognized responsibility has been shared by manufacturers, physicians, 
and motion picture technicians to simplify the technic and improve 
the quality of clinical and surgical records to be provided by motion 
pictures. 

The special application of telephoto lenses, the development of 
faster emulsions, improvements in the design of cameras, and the 
more efficient lighting are available advantages. All these are 
familiar expressions of advancements in this field of medicine. 

The application of Kodacolor Film in medical photography was 
thoroughly studied by the writers a number of years ago. Satis- 
factory color values could be obtained although certain limitations 
prevailed. Among these it was evident that quality of the pictures 
was defined by the focal length of available lenses, the amount of 
auxiliary illumination was a detriment in the operating room, while 
the necessity for using filters, both in making and projecting the 



* Presented at the Fall, 1936, Meeting at Rochester, N. Y. 
** School of Medicine, University of Rochester, Rochester, N. Y. 
t Eastman Kodak Company, Rochester, N. Y. 

173 



174 R. P. SCHWARTZ AND H. B. TUTTLE [J. S. M. P. E. 

films, further restricted the usefulness of this process of motion pic- 
tures in the field of medicine. 

All these limitations are now removed. The new artificial- 
light type of Kodachrome records the actual color on the film. By 
virtue of new lenses, minimum accessory lighting, and reproduction 
of color values with precision, the ideal film is now available for all 
the fields of medicine in which records can be improved by the per- 
manent and lasting reproduction of color. Limitations heretofore 
expressed regarding the application of black-and-white or Kodacolor 
film to medical photography no longer hold with regard to the new 
artificial-light Kodachrome Film. 

With the addition of color, it is possible to bring to the motion 
picture screen all the realism of the operating room, thus enabling 
medical students to see not only more of the operations but to see 
the surgical field in the same size and in the same perspective as 
the surgeon sees it. Unusual cases can not only be projected many 
times for a given group, but can be used for teaching and lecturing, 
and as general case-history records over periods of many years. 

Until recently it was necessary to use a blue artificial light com- 
pensating filter when making a medical film with Kodachrome 
Film. Daylight Kodachrome Film when used with the blue filter 
rendered satisfactory color-films; but the filter, which had an ex- 
posure factor of four, or two diaphragm stops, necessitated the use 
of large-aperture lenses or a degree of illumination of the surgical 
field that was impracticable. 

The artificial-light Type A Kodachrome Film eliminates many of 
the disadvantages of the past. No filter is necessary with it. Its 
high speed has made it possible to use regular operating room lighting 
equipment, and, because of the high speed, all standard types of 
telephoto lenses can be used. 

Since the film is color-balanced specifically to match artificial or 
tungsten light, it is obvious that if it were used in daylight, the 
high blue-sensitivity of the film would cause a predominance of blues. 
Therefore, care must be exercised to eliminate daylight from the 
room when it is used. The film can, of course, be used in sunlight 
by using an orange filter, which is complementary to the blue filter 
formerly used with daylight Kodachrome and artificial light. 

It has been found that if any standard type 500-watt spotlight 
is added to the regular operating room surgical field illumination, 
a lens aperture of //5.6 can be used. This arrangement is particu- 



Feb., 1937] MEDICAL MOTION PICTURES IN COLOR 175 

larly desirable when the 2-, 3-, 4Vr, or 6-inch telephoto lenses are 
employed. 

The surgical field should be lighted with a flat, even illumination. 
The 500-watt auxiliary spot and regular operating room lamps 
should be directed from above, so that none of the light is cut off 
from the surgical field by the operating team. The 2-, 3-, 4Va- 
and 6-inch telephoto lenses enable a camera technician to work at a 
respectful distance from the operating table, thus eliminating the 
hazard of contaminating the sterile field. 

Color motion pictures are actually easier to make than black- 
and-white. The addition of color more than justifies the additional 
cost of the film. The medical student now can see not only surgical 
technic and general detail, but all the slight variations in color of 
skin and tissue, and the general pathology of the case as well. 

Gross specimens can be filmed in color and the pictures preserved 
in gelatin at a lower cost than it is possible to preserve the specimens 
in alcohol or formaldehyde. Furthermore, the true colors of the 
fresh specimens will be preserved in films, whereas the specimens 
change color when preserved in jars. Medical films in color are not 
limited in use to medical schools and hospitals ; recently medical 
color-films were used as evidence in deciding a court action. The 
possibilities of using properly designed films in grade schools for 
teaching hygiene have not yet been explored. 

One has only to pause a moment and imagine the value films would 
have today if it had been possible to record in this fashion the ex- 
periments of Pasteur, the first use of anesthetics, or the technic of any 
of the early surgeons who made medical history. 

There is perhaps more history-making being done today than ever 
before in medical science. In many instances color motion picture 
records are being made of these new and startling history-making 
events. It is hoped that in a short time every teaching hospital 
and medical school will be properly equipped to portray in color- 
films, not only to the medical student, but to the layman as well, 
the fruits of their important discoveries. 

(-4 motion picture of medical subjects in color made on the Type A film was projected 
at the close of the paper.) 

DISCUSSION 

MR. TUTTLE: When making the picture, the field was illuminated with one 
medical spotlamp, and exposures were made with an aperture of //4, on the arti- 
ficial-light type of film that has been available for the past few months. Within 



176 R. P. SCHWARTZ AND H. B. TUTTLE [J. S. M. p. E. 

the past week or so, new fast artificial-light film has been placed upon the market 
that is twice as fast as the old film, so that the same scene can be filmed today at 
//5.6 with one medical spotlight equipped with a 500-watt, 100- to 105-volt lamp. 

This particular surgery was filmed with a regular 1-inch lens with an auxiliary 
lens fastened over the front of it. The same field size could be obtained by using 
the 3-, 4Y 2 -, or 6-inch lens with the focusing set screws removed. A 5-diopter 
lens was used, with the subject 12 to 14 inches from the camera. The danger 
of contaminating the field of surgery is not as great in eye operations as in ab- 
dominal or other operations, so that the camera can be placed quite close. 

MR. GILMOUR: A student recently expressed himself to me in this manner: 
"I have seen operations in amphitheaters, and being young, and, perhaps very 
human, I was constantly mentally upset as to whether the operation would turn 
out successfully or not. It happens that most of my work is in obstetrics, and I 
worry as to whether the deliveries are going to be successful. On the other hand, 
when I am shown a picture, I know that the operation has been a successful one 
and that the technic was perfect. I am not worried, and I get a great deal more 
out of it as a film than as an operation." 

Another point has to do, not necessarily with medical color photography, but 
with any type of color photography, as to why we feel that the color upon the 
screen seems more brilliant than the color we see in the original. The thought 
that I have is that when we darken the room for projection, the color seems very 
vivid by contrast with the darkened room. 

MR. TUTTLE : As regards the first question, I think it is generally accepted as 
a fact by most doctors in teaching hospitals, that students get very little out of 
sitting in the gallery and watching operations. It is impossible for the master 
surgeon to impart very much to them because his responsibility to the patient is 
uppermost in his mind. Usually the student has to stay well back, and can not 
see the things that the doctor can see because of the angle and perspective of 
observation. 

Motion pictures are ideally suited for teaching surgery because after the opera- 
tion is over the surgeon who performed it can go into the teaching room and pro- 
ject the film any number of times. He can prepare a lecture to fit the picture, or 
cut the picture to fit his lecture, as the requirement may be, to cover the case more 
completely. 

The second question probably has been considered by everyone interested in 
color photography. The psychological effect of seeing an original subject, and 
then going into a darkened room and seeing a picture of it projected upon the 
screen gives one quite a difference in the personal interpretation of color. Ima- 
gine, for instance, one who has never been, say, to Bryce Canyon. He gets off 
the train, the sun shining, the grass green, the flowers filling the air with their 
aroma. The birds are singing. He has received good news from home, and his 
spirits are high. All his other senses are contributing to his visual sense. After 
looking everything over and seeing the beautiful colors, he decides to take a movie 
of it. 

He puts the camera to his eye. It has a limited angle of view, and must record 
with only one "eye." A picture is made that we shall assume, for all practical 
purposes, is perfect. After development, processing, and so forth, it comes out 
as nearly technically perfect as it is possible to make a colored picture. 



Feb., 1937] MEDICAL MOTION PICTURES IN COLOR 177 

Returning from his vacation to a pile of work upon his desk, he finds the good 
news from home was perhaps not so good. His spirits are lowered; perhaps he 
has a headache. He goes into a darkened room where he sees only a few degrees 
of the entire picture. He remembers seeing the scene of course, over a much 
greater angle, and the results are very disappointing because all the essential 
effects of the other senses are missing or are different. 

However, on the other hand, if he should go to Bryce Canyon with one of his 
eyes bandaged, and his ears and nose plugged up so that those senses are not 
affected; and sees only what the camera sees through its limited finder; and sub- 
sequently goes right back without seeing any more, has the film processed, and 
looks at it in the darkened room, he is amazed at the beautiful colors he has re- 
corded. 

MR. CLARK: Kodachrome Type A is balanced for illumination of Photoflood 
quality. Can not the fact that you make the pictures with normal tungsten il- 
lumination account for a certain increase in the redness or yellowness of the pic- 
ture? 

MR. TUTTLE: The pictures were not made with regular tungsten. The lamp 
was slightly overvolted. A 500-watt lamp, 100 to 105 volts, was used on a 115- 
to 118-volt line. 



NEW METHOD FOR THE DRY HYPERSENSITIZATION 
OF PHOTOGRAPHIC EMULSIONS* 



F. DERSCH AND H. DURR** 



Summary. Hypersensitization by mercury vapor increases the speed of photo- 
graphic negative emulsions about 50 to 150 per cent, depending upon the emulsions 
used for the treatment. The important features of this method that make it superior 
to the well known wet-hyper sensitizing methods are: 

(1) The film does not have to be put through a bathing process and then dried. 
(2) The mercury vapors are active also upon tightly wound spools of film, the sensitiz- 
ing effect being uniformly spread over the whole length (e. g. t of a 1000-foot roll of 35- 
mm. motion picture film). If sufficient time is available for hyper sensitizing, the 
films need not even be removed from their original wrappers, as the mercury vapors 
diffuse sufficiently through the wrapping material. (5) The increase of sensitivity is 
general throughout the range of wavelength of light to which the film was originally 
sensitive. (4) Not only can unexposed film be hype%sensitized by this method, but it 
is also possible to intensify the latent image with mercury vapors. (5) The stability 
of the film is not permanently affected, although the increase in speed is gradually lost 
over a period of four weeks of aging. The clearness, however, remains the same, and 
may even improve somewhat. By a second treatment with mercury vapor the hyper- 
sensitization can be renewed in a film that has recovered from previous hyper sensitiz- 
ing. 

After the introduction of panchromatic emulsions, methods of in- 
creasing the sensitivity of these emulsions by special treatments be- 
came generally known by the name of "hypersensitization." These 
methods were based upon the well known fact that the sensitivity of 
photographic films and plates can be increased by bathing them in 
water or in solutions containing small amounts of ammonia. Later, 
other solutions were recommended for the purpose; for example, 
solutions containing small amounts of a silver salt and ammonia in 
water, or solutions containing small amounts of silver nitrate and hy- 
drogen peroxide, and so on. The increase of speed attainable with 
this weMrypersensitizing method, as it might be called, amounts to 



* Received October 15, 1936; presented at the Fall, 1936, Meeting at Roches- 
ter, N. Y. 

** Agfa Ansco Corp., Binghamton, N. Y. 

178 



HYPERSENSITIZATION OF EMULSIONS 179 

100 per cent, more or less, depending upon the type of emulsion used. 
Emulsions that have been made in the presence of ammonia usually 
show less increase of speed. 

To make the special treatment practicable, for instance, with pan- 
chromatic cine negative film, great care had to be observed in manipu- 
lating the wet films, and redrying the emulsion carefully was particu- 
larly important to the quality of the results. In addition, hyper- 
sensitized films always have certain disadvantages, especially with 
regard to their keeping qualities, which will be discussed later. 

With the introduction of the supersensitive types of negative film, 
interest in these inconvenient, cumbersome, and expensive methods 
of hypersensitization declined considerably. 

Yet, even with the availability of the supersensitive types of pan- 
chromatic materials there still exists, and probably always will, a 
demand for higher sensitivity, if possible without increasing the 
graininess, especially for processes in which only a rather unsatisfac- 
tory compromise can be attained between illumination and the most 
desirable lens aperture. 

From this point of vieW, it appears that the method described in 
this paper may prove to be not only interesting from the theoretical 
standpoint, but will also offer prospects of its practical application. 

GENERAL 

Upon investigating the effect of mercury vapor upon photographic 
emulsions in the Agfa Research Laboratories, it was found that the 
sensitivity of nearly all types of negative emulsions can be consider- 
ably increased when dry films or plates are exposed to the action of 
mercury vapor. 

This effect was first noticed in 1928 by Baukloh, 1 who found that 
mercury vapors sensitize silver bromide and, more especially, silver 
iodide emulsions. Apparently nothing has been done since to ac- 
count for this effect in theoretical considerations or to make practical 
use of it. Investigations by Winther 2 in 1933 even contradict in part 
the findings of Baukloh. 

The action of the mercury vapor was found to be relatively slow. 
In our original experiments, photographic emulsions upon films 
and plates were exposed to the action of mercury vapor by placing 
them into a light-tight container, the bottom of which was covered 
with a thin layer of metallic mercury. In the container, films and 
plates were treated for approximately thirty hours, after which the 



180 



F. DERSCH AND H. DURR 



[J. S. M. P. E. 



Agfa Cine Negatir* 
Superpan 



- Not Treated 
B - Dry-Hypersensltized 
with Meroury Vapbr 



Agfa Cine Negative 
Superpan 



A. - Not Treated 

B - Dry-Hyperaenaitiaed with 

Meroxiry' Vapor 
C - W-et-Hypersensitized with 



Relative Xrpoure 

I 

126 



Agfa Cine Negative 
Superpan 



A - Not Treated 

B - Emulsion Treated with Mercury 

Vapor before Exposure 
- Emulsion Treated with Mercury 

Vapor after Exposure 







0.2 . 



256 



FIG. 1. (Upper) Difference of sensitivity between untreated 
material and material hypersensitized by mercury. 

FIG. 2. (Center) Difference of sensitivity between material treated 
with mercury vapor and material hypersensitized by a wet -hyper- 
sensitizing method. 

FIG. 3. (Lower) Effect of mercury hypersensitization before and 
after exposure. 



Feb., 1937] HYPERSENSITIZATION OF EMULSIONS 181 

emulsions showed an increase of sensitivity of about 75 to 150 per 
cent, depending upon the type of emulsion and upon the mercury 
vapor concentration within the container. 

In Fig. 1 the difference of sensitivity between the untreated mate- 
rial and the material hypersensitized by mercury can be seen, and in 
this particular case is about 75 per cent. It is interesting to note that 
the characteristic curve of the mercury-hypersensitized emulsion 
runs almost parallel to the curve of the untreated material. This 
fact is pertinent because wet-hypersensitized materials usually show 
a distinctly steeper gradient than the untreated materials, as will be 
seen in Fig. 2. 

In Fig. 2 the characteristic curve of the untreated material, A, is 
plotted together with the curve of the same emulsion treated with 
mercury vapor, B. The third curve, C, is for the same emulsion 
again, but hypersensitized by one of the wet-hypersensitizing meth- 
ods. In this case a small amount of ammonia in distilled water was 
used as the hypersensitizing solution. In Fig. 2 the wet-hypersen- 
sitized emulsion shows a somewhat steeper gradient than either the 
type emulsion or the emulsion dry-hypersensitized by mercury vapor. 
This increase of gamma is characteristic of wet-hypersensitizing 
methods, while the hypersensitizing by mercury vapor has prac- 
tically no influence upon the gradient as far as the useful part of the 
curve is concerned. 

The increase of gamma of wet-hypersensitized panchromatic emul- 
sions is largely due to the fact that bathing methods increase the sen- 
sitivity of panchromatic emulsions in the yellowish green and red- 
sensitive portions of the spectrum much more than they do in the blue. 
The original ratio of sensitivity, for instance, the blue-yellow or blue- 
red ratio, becomes changed, which means that the filter-factors of 
the wet-hypersensitized emulsions are different from those of the 
original emulsion. In this respect the dry-hypersensitized film be- 
haves in a different manner. The mercury does not change the origi- 
nal sensitivity ratio in different wavelength regions; it appears that 
the increase of sensitivity is proportional throughout the portions of 
the spectrum to which the emulsion was originally sensitive. This 
method of dry-hypersensitizing apparently does not change the filter- 
factors of the original emulsion. 

Film and plate emulsions from various manufacturers have been 
treated with mercury vapor, and no fundamental differences in be- 
havior could be found. There is also no significant difference be- 



182 F. DERSCH AND H. DI)RR [J. S. M. P. E. 

tween the effect of mercury vapor upon ammonia and upon non- 
ammonia types of emulsions. It has been mentioned already that the 
action of mercury vapor is rather slow. At normal room tempera- 
tures, unwrapped films must be exposed to the vapors for at least 
twenty-four to thirty hours before the maximum increase of speed is 
attained. Longer treatment with mercury vapor does not increase 
the sensitivity to an appreciable extent, but the fog gradually in- 
creases. It would, of course, not always be practicable to treat un- 
wrapped and unrolled films for thirty hours in an atmosphere contain- 
ing mercury vapor. However, it has been found that it is not at all 
necessary to unwind and unwrap the films completely. The penetra- 
tion of the mercury vapor into spooled and tightly rolled material is 
surprisingly uniform and efficient, making the whole process much 
more practicable and convenient. It is, for instance, sufficient to 
leave a 1000-ft. roll of motion picture negative film in the original can, 
and put a few drops of mercury wrapped in porous paper inside the 
empty space of the film core. The film can must be closed and sealed 
with tape, and should stand for approximately six to eight days. 
During that period an increase of speed extending very uniformly 
throughout the entire 1000-ft. roll can be noticed. The same effect 
can, of course, be attained with regular rollfilm spools or with spools 
for the Leica and Contax cameras. In the latter case it is not neces- 
sary to open the original cartridge; it is sufficient to put the whole 
cartridge into a small container containing mercury. 

STABILITY OF DRY-HYPERSENSITIZING 

The hypersensitization effected with mercury vapor is not perma- 
nent. The speed gradually recedes over a period of about four 
weeks; after which a more or less stable condition is reached when the 
sensitivity of the material is somewhat below that of the emulsion 
before the treatment. However, during the aging period, the dry- 
hypersensitized emulsion remains free from fog. After three to four 
weeks the fog value of the emulsion is even somewhat lower than the 
fog value of the original film. It is known that the stability of films 
or plates that have been hypersensitized by bathing methods is very 
poor. The fog of the emulsion rapidly increases with age, and 
materials so treated are usually ruined by excessively high fog in 
about four weeks. This is another distinct difference in behavior 
between wet- and dry-hypersensitized materials. After losing their 
additional sensitivity, dry-hypersensitized emulsions are still in a 



Feb.. 1937] HYPERSENSITIZATION OF EMULSIONS 183 

usable condition. The speed is somewhat less than that of the origi- 
nal untreated film, but the clearness is at least the same or better. 
There is another advantage. Emulsions that have been treated with 
mercury vapor, but have not been used before losing the additional 
sensitivity, can be re-hypersensitized by treatment in the mercury 
atmosphere a second or even a third time. 

As far as could be seen, by comparing treated material with an un- 
treated type, the grain size was not noticeably affected. 

EFFECT OF MERCURY VAPOR UPON THE LATENT IMAGE 

So far only the effect of mercury vapor upon unexposed photo- 
graphic emulsions has been considered. Theoretical considerations, 
which will be discussed later, led to the discovery that the effect of 
mercury vapor upon the latent image is even greater than it is upon 
the unexposed emulsion. This action may probably be better de- 
scribed by the expression intensification of the latent image, as it has 
been applied to similar processes utilizing hydrogen peroxide. 

In Fig. 3, A is the characteristic curve of an untreated emulsion; B 
is for the same emulsion dry-hypersensitized by mercury before ex- 
posure; and C is for the same emulsion, but in this case the mercury 
vapor treatment took place after exposure in other words, the 
latent image has been intensified after exposure but before develop- 
ment. From the curves it can be seen that the effect of the mercury 
upon the latent image is distinctly greater than it is upon the unex- 
posed emulsion. However, except for the difference in intensity of 
the effect, the characteristic behavior is in both cases the same. The 
characteristic curve of the intensified latent image, as can be seen in 
Fig. 3, also runs almost parallel to the original characteristic curve. 
The stability of intensification of the latent image is limited as to the 
length of time between treatment and development, as is the hyper- 
sensitization of the unexposed emulsion as to time between treatment 
and use. 

The treatment of the exposed film with mercury vapor to intensify 
the latent image can be done exactly in the same manner as has been 
described for dry hypersensitization. It is, therefore, possible to 
correct an underexposed picture by treating the undeveloped film 
with mercury vapor for a certain length of time, provided, of course, 
underexposure is known or suspected. After the treatment, the film 
is developed as usual, and will produce a negative similar to one ex- 
posed with 100 to 150 per cent more light. Tightly wound rolls in 






184 F. DERSCH AND H. DURR [J. S. M. P. E. 

cans can be hypersensitized if sufficient time, generally six to eight 
days, is allowed. Due to the relatively slow action, good penetration 
to all the layers of emulsion is achieved, and the effect is more or less 
uniform throughout the roll. 

PRACTICAL APPLICATIONS 

Within the scope of this paper it is possible to describe only very 
briefly how the material should be handled to obtain the best results. 
As a matter of fact, it would be very difficult to give exact formulas. 
Fortunately it is not necessary to do so ; because of the slowness of the 
effect, the time of treatment, and the mercury vapor concentration 
do not have to be very exact. As a general rule, loose and unwrapped 
material should be treated from 30 to 40 hours at room temperature, 
while wrapped and spooled materials require treatment for seven to 
ten days, in a mercury vapor concentration created, for instance, by 
0.5 gram of mercury in a 1000-ft. film container. In place of liquid 
mercury, of course, all compounds, amalgams, such as silver amal- 
gams, and other preparations that emit mercury vapors can be used 
for dry hypersensitization or for intensifying the latent image. 

THEORETICAL 

In attempting to coordinate the mercury vapor effect with existing 
theories on the photographic sensitivity of silver halide emulsions and 
on the latent image, it is necessary to mention briefly, at least, these 
contemporary theories first. According to prevailing concepts, the 
sensitivity of silver halide gelatin emulsions is attributed mainly to 
the presence of silver and silver sulfide nuclei upon or within the 
silver halide grain. Theories proposing the existence of these me- 
tallic silver nuclei, so-called photo-silver, as well as the presence of 
silver sulfide, have been quite generally accepted. Both types 
of nuclei are already present on the unexposed silver halide grain. 
The number of atoms of silver contained in one of these nuclei has 
not yet been definitely agreed upon, and the nuclei can not be seen 
on unexposed silver halide grains, even with powerful microscopes. 
The inter-reaction between the silver sulfide specks and the photo- 
silver nuclei seems to be quite complicated, and no entirely satis- 
factory theory of the mechanism of this inter-reaction with re- 
gard to the sensitivity of the silver halide grain has been advanced. 
It might be possible that the silver sulfide specks activate or concen- 
trate the relatively few photo-silver atoms upon the grain, as sug- 









Feb., 1937] HYPERSENSITIZATION OF EMULSIONS 185 

gested by Sheppard, 3 or that silver sulfide disturbs the crystal struc- 
ture and thereby facilitates the formation of larger silver nuclei under 
the influence of light. 4 The effect of mercury upon photographic silver 
halide emulsions seems to be more closely related to the presence of 
metallic silver nuclei or photo-silver on the grains of ripened silver 
halide emulsion, rather than to the presence of the silver sulfide. Ac- 
cording to the present theories of the latent image, the metallic silver 
nuclei are of great importance to the developability of silver halide 
grains. This photo-silver is supposedly already present on ripened, 
but unexposed, silver halide grains. However, these nuclei are not 
yet large enough to catalyze the action of the developer upon any 
particular grain. If the silver nuclei become larger, through the 
formation of more photo-silver by exposure or by prolonged ripening, 
the size of these nuclei, or the number of silver atoms, reaches or 
passes a certain critical point at, or beyond, which the nuclei can suf- 
ficiently catalyze, or accelerate, the reaction of the developer, and 
the grain thereby is rendered developable. This theory of the action 
of the silver nuclei and of the latent image has been very helpful in 
describing the peculiarities of the mercury vapor effect. It can be 
assumed that the silver nuclei absorb or adsorb mercury atoms, and 
possibly form silver amalgams. The size of the silver nuclei will be- 
come increased by these mercury atoms, which, in turn, could ac- 
count for an increase of sensitivity of any particular grain. In case 
too much mercury is absorbed, the grain becomes developable with- 
out exposure, causing the emulsion to show fog. 

Following this general line of thought, it might be expected that 
the mercury vapor would show a greater effect upon the latent image 
than upon unexposed material. Experiments in this direction con- 
firmed the theoretical speculations. The exposed silver halide grains 
contain a much greater amount of silver as nuclei, and the chances of 
mercury atoms being absorbed by those exposed grains are that much 
greater. The gradual loss of hypersensitization upon aging might be 
due to the loss of mercury through re- vaporization. The final stage 
at which the sensitivity of the treated and aged emulsion arrives and 
at which it apparently remains, wherein the speed is somewhat less 
than the speed of the original untreated material, might be attributed 
to the possible formation of mercury salts. 

These explanations for the mercury effect are, of course, still hypo- 
thetical, but the effect apparently opens a new and very interesting 
avenue of approach to the theory of the latent image. 






186 F. DERSCH AND H. DURR [J. S. M. P. E. 

REFERENCES 

1 Zeitschr.fur wids Phot., 25 (1928), p. 250. 

2 Ibid., 32 (1933), p. 157. 

3 SHEPPARD, S. E. : Third Colloidal Symposiums (monograph). 

4 NEBLBTTE, C. B.: "Photography, Its Principles and Practices," McGraw- 
Hill Pub. Co., (1930), New York, p. 207. 

DISCUSSION 

MR. CRABTREE: After storing the film in the presence of mercury, have you 
made uniformity tests by flashing and developing the film on a machine? 

MR. FYFE: Yes. The variation is no more than 25 per cent either way. 

MR. CRABTREE : It is well known that you may get effective hypersensitization 
or growth of the latent image by merely storing the latent image. Are these hy- 
persensitization values over and above the latent image growth on keeping? 

MR. SCHMIDT: I might mention that the work was done in two different ways, 
practically and sensitometrically. Film was actually used in the camera, part of 
it exposed, and hypersensitized. During hypersensitization the latent image was 
kept on parts that were not given the treatment, and both parts were developed 
together. 

MR. SHORNEY: What effect would hypersensitization have upon Kodachrome? 
Would it affect only the upper emulsion, or would it penetrate to the lower emul- 
sions? 

MR. FYFE: I can not answer that. We have not investigated it. 

MR. MATTHEWS: I assume that the phenomenon is so new that the investiga- 
tions have been restricted largely to existing high-speed emulsions. 

MR. SCHMIDT: We tried different emulsions, including process and x-ray emul- 
sions. It was interesting to note that with x-ray emulsions the image produced 
by x-rays and by the short-wave and ultraviolet rays of the fluorescent screen 
showed intensification. 

MR. COOK: Is there a detectable amount of mercury in the film after absorp- 
tion? 

MR. FYFE : Most likely not. The mercury vapor concentration was the con- 
centration at room temperature, and was very, very low. The amount taken up 
by the film is probably immeasurable by ordinary methods. 

MR. CRABTREE: What are the effects of temperature and of storing at pres- 
sures below atmospheric? 

MR. FYFE: We did not extensively investigate the effect of temperature; we 
worked at room temperatures. I did make some experiments at lower pressures 
in vacuum, and apparently there was no change of effect. We got the same ef- 
fect at reduced pressures as we did at normal atmospheric. 

MR. CRABTREE: What is the concentration of mercury in the vapor phase at, 
let us say, atmospheric pressure and a temperature at 70 degrees? 

MR. SCHMIDT: I do not know exactly what the vapor pressure is; 0.0017 milli- 
meter, I think. The percentage of mercury is very low. We are not interested 
in increasing the amount. We would rather work in the other direction. We 
tried to have definitely lower pressures because that seems to be more promising. 



Feb., 1937] HYPERSENSITIZATION OF EMULSIONS 187 

Evidently, the number of mercury atoms present is much greater than is neces 
sary to show an effect. 

MR. CRABTREE: What is the critical mercury concentration necessary? 

MR. SCHMIDT: We believe there is a lower concentration that produces an op- 
timal effect. It is undesirable to have excessive amounts of mercury vapors in 
photographic laboratories. 

MR. MISENER: Would there be any difference in the hypersensitization if the 
latent image were treated immediately after exposure or after a keeping period of a 
few weeks? 

MR. FYFE: As I remember, there was no change in the effect, but we have not 
gone deeply enough into that to answer definitely. 

MR. MILLER: Was any attempt made to accelerate the effects and produce 
them in a shorter length of time, such as by raising the temperature? 

MR. FYFE: The temperature was increased in one or two experiments; and 
since we ran into fog troubles, and since mercury vapor concentrations at tempera- 
tures higher than room temperature are not desirable, we did not investigate fur- 
ther in that direction. 



REPORT OF THE MEMBERSHIP AND SUBSCRIPTION 
COMMITTEE* 



The growth of the Society continues at a satisfactory pace. Two 
hundred and thirteen new members have been added to the rolls 
this year, ten old members have been reinstated, and, in addition, 
twenty-five applications are pending. As is usual each year, a number 
of members have allowed their memberships to lapse, and several 
resignations and deaths have occurred. However, despite these losses, 
the membership now totals 950 domestic and 325 foreign members, 
or a grand total of 1275. 

While this number is a fairly impressive one, the Society is not 
serving as many individuals as it should. Many engineers, tech- 
nicians, and others prominent in the industry are still not affiliated 
with the Society. Our task is to convince them of the advantages of 
membership in the Society, and the Committee requests the assistance 
of the members in this work. 

The charts accompanying this report show the distribution of 
membership throughout the various states of the United States, 
and throughout the countries of the world. It will be observed that 
37 states and 33 different countries are represented. New York 
State leads with 365 members. California follows with 180 members. 
Unfortunately, the distribution by localities is not well shown by the 
chart. For example, quite a large group of those who are active in 
the New York district reside in Pennsylvania and New Jersey ; and 
many of those in the locality comprising Philadelphia and Camden 
reside in the two states, New Jersey and Pennsylvania. 

As for the distribution of membership according to the three local 
sections of the Society, viz., Atlantic Coast, Mid- West, and Pacific 
Coast, this distribution is determined by a division of the country 
into three sections by parallels of longitude drawn through points 
forty miles west of Cleveland and forty miles west of Denver. 

Non-member subscriptions for the JOURNAL are likewise increasing, 
although at a much slower rate because the Society is primarily a 

* Presented at the Fall, 1936, Meeting at Rochester, N. Y. 
188 



MEMBERSHIP AND SUBSCRIPTION 



189 




1916 1920 1924 19Z8 1932 1936 

FIG. 1. Growth of membership. 


































msm 


BBHH 



i 




mm 
















i 



B 

1 


















Doa 


oesti 


> Mam 


jersh 


LP 


1 




















Society of 






















Motion Picture 


[ 




















Engineers 
























I! 


)36 









a 


o 


1C 


H> 


1! 


O 


2C 


O 


2 


50 


* 


M> 


3! 


O 



W. Va. 
Wis. 



FIG. 2. Domestic membership. 



190 



MEMBERSHIP AND SUBSCRIPTION 



membership organization. During the present year there have 
been 111 new subscriptions and 65 expirations, making a net increase 
of 46. During the past two years the number of subscriptions has 
increased by nearly 40 per cent. These subscriptions are taken mostly 
by public libraries and by the libraries of educational and large in- 
dustrial institutions. 

E. R. GEIB, Chairman 



Argentina 

Australia 

Austria 

Bermuda 

Brazil 

Canada 

Canal Zone 

China 

Cuba 

Czechoslovakia 

England 

France 

Germany 

Hawaii 

Holland 

Hungary 



Mexico 

New Zealand 

Norway 

Philippine Isl. 

Poland 

Portugal 

Roumania 

Russia 

Scotland 

Spain 

Switzerland 
Wales 



a 

















mmm 

1 


i 






1 
















BBi 

mam 


MM 
















mam 







i 






Tor 


>ign 


ktembe 


>ship 










Society of 








Motion Picture 


I 






Engineers 




1. 








1 


)36 






i 


2 


3 


4 


o e 


6 


"J 









FIG. 3. Foreign Membership. 



NEW MOTION PICTURE APPARATUS 

During the Conventions of the Society, symposiums on new motion picture appara- 
tus are held, in which various manufacturers of equipment describe and demon- 
strate their new products and developments. Some of this equipment is described in 
the following pages; the remainder will be published in subsequent issues of the 
Journal. 

A NEW HIGH-QUALITY PORTABLE FILM-RECORDING SYSTEM* 
F. L. HOPPER, E. C. MANDERFELD, AND R. R. SCOVILLE** 

In addition to providing sound records of high quality, a modern recording sys- 
tem must be compact in size, mobile, and require a minimum in the way of operat- 
ing technic and maintenance. The older type of bulky fixed channel is rapidly 
being displaced by the newer units. 

The new portable film-recording channel described in this paper has been de- 
veloped as a result of experience in the field with a previous portable channel, 1 and 
is suitable both from a quality and operating standpoint for studio or location 
work, simplifies maintenance, and makes possible a single operating technic. 

A complete channel consists of: a pick-up or stage unit, containing mixing and 
monitoring facilities; a main amplifier supplying the light -valve and the monitor- 
ing system ; a noise-reduction unit of the carrier type, providing biasing current 
for the light-valve; a high-quality film-recorder, with associated motor system 
and recorder control unit ; and a gain set and oscillator providing transmission- 
testing facilities. Power for the entire channel may be supplied either by an a-c. 
operated unit or by batteries. 

All units are interconnected by means of six conductor cords equipped with 
jacks and plugs. A schematic drawing showing the equipment connected for 
operation is shown in Fig. 1. The units comprising the recording channel are 
shown in Fig. 2. 

General Design Features. In the design of the system advantage has been taken 
of recent developments by Bell Telephone Laboratories in the fields of vacuum 
tubes, 2 communication transformers, 3 and moving-coil microphones 4 and head- 
sets. Contacts with the studios have established operating requirements and the 
need for certain adjuncts to the sound-recording system to enhance its useful- 
ness. Consideration of these factors has resulted in a recording system that is 
reliable, and easily operated and maintained. 

* Received October 10, 1936; presented at the Fall, 1936, Meeting at Roches- 
ter, N. Y. 

** Electrical Research Products, Inc., New York, N. Y. 

191 



192 



NEW MOTION PICTURE APPARATUS [J. S. M. p. E. 



Each unit is of a new style of construction designed to make the equipment more 
compact, render the parts readily accessible, and provide adequate protection 
against rough handling in the field, and results in a considerable saving of space 
and weight. All units use the chassis type of construction, consisting of a com- 
bination of vertical and horizontal panels, rigidly braced and carrying the equip- 
ment and controls. All the apparatus is mounted on one side of the panel, the 
wiring being done on the opposite side, allowing short, direct connections be- 
tween the various circuit elements. A jack panel is located at the rear of the unit, 
affording easy access to patching facilities. All panels carrying controls are 
covered with a thin mat which carries the engraving and is put on after the unit 
has been assembled, thereby assuring freedom from injury during construction. 
Fig. 3 shows the panel of a typical unit. Figs. 4 and 5 show, respectively, the 



MICROPHONES 




FIG. 1. Schematic arrangement of channel. 



equipment and wiring sides of the same unit. The assembled chassis may be slid 
into a case for general use, or may be easily adapted for mounting in mixer tables, 6 
or upon relay racks. 

Shock-absorbing mountings are employed for all vacuum tubes, consisting of 
a channel carrying the tube sockets, the channel being fastened to the chassis 
through isolating rubber mountings. This form of support is compact and quite 
effective. Shielding has been employed between vacuum tubes and equipment 
parts wherever necessary to eliminate cross-talk. Low-level transformers are 
also shielded magnetically with permalloy to reduce pick-up due to exposure to 
external power fields. All filament and plate circuit metering is accomplished by 
means of a switch and meter. The meter indicates percentages, reading 100 per 
cent when operation is normal. 

Transmission Features. Heater type vacuum tubes are employed throughout 
the recording system, which allows operating the channel on either a-c. or d-c. 
sources. Special tubes have been employed wherever required by the particular 
functions they are called upon to fulfill. Tubes having extremely low noise and 
microphonic outputs are used in the low-level stages. All tubes are self -biased, re- 
quiring no C batteries. Alternating-current operation may be effected by one of two 



Feb., 1937] 



NEW MOTION PICTURE APPARATUS 



193 



methods: The first requires a power unit supplying rectified current of the A and B 
circuits, and is normally supplied as a part of the channel. The second con- 
sists in supplying alternating current directly to all the heaters by means of a small 
filament transformer, and direct current to the plates from a rectifier. It is pref- 
erable from noise considerations to supply the heaters of the pick-up unit with 
direct current. Noise measurements indicate that such a-c. operated channels 
are practically as quiet as those operated completely by batteries. 

The grounding system for the channel has been designed to provide protection 
from exposure to both radio and power circuits. The speech circuit ground is 
carried back from the pick-up unit to the main amplifier separately, and then con- 
nected to the system ground. All shields from the microphone cables, wiring in 
the pick-up unit, and shields in the transmission and signal cables are connected 




FIG. 2. Entire channel set-up. 



through by means of a conductor to the system ground at the main amplifier. 
All grids and plates are protected by filtering, resulting in a system free from vari- 
ous forms of feedback. The maximum overall system gain is 110 db., which is 
about equally divided between the pick-up unit and the main amplifier. The 
frequency-response characteristic is essentially uniform from 40 to 10,000 cps. 
The amplifier system output is adequate for light-valve operation. 

Pick-Up Unit. The pick-up unit is contained in a case measuring 18 X 10 X 
10 3 /4 inches, and weighing 47 pounds. It contains a four-position mixer, bal- 
anced to ground. Three mixer positions are supplied from built-in microphone 
amplifier stages. Provision has been made for connecting a fourth microphone 
by means of an external microphone amplifier. Microphone cut-off keys are as- 
sociated with each mixer position. The output of the mixer is further amplified 
by a built-in two-stage booster amplifier containing a variable dialog equalizer and 



194 



NEW MOTION PICTURE APPARATUS [J. S. M. P. E. 




FIG. 3. (Upper) Pick-up unit. 

FIG. 4. (Center) Equipment side of interior of pick-up unit. 
FIG. 5. (Lower) Wiring side of interior of pick-up unit. 



Feb., 1937] NEW MOTION PICTURE APPARATUS 195 

an inclusive volume control. The normal gain of the unit is approximately 55 db., 
with a substantially uniform response from 40 to 10,000 cps. An internal gain 
adjustment has been provided so that the gain may be increased to 64 db. to meet 
special working conditions. Cross-talk between mixer stages has been eliminated 
by means of careful wiring and shielding. 

A power level indicator of the high-speed type is employed for mixing. It is 
bridged across the output stage, supplying the light-valve, and under operating 
conditions contributes no bridging loss. A sensitivity potentiometer allows ade- 
quate adjustment of its range. 

Monitoring is accomplished by means of a high-quality moving-coil head-set. 
Means are provided for switching the head-set from direct monitoring to photo- 
electric cell monitoring, a volume control common to both circuits allowing ad- 
justment of volume in the head-set. 

The order-wire system is designed to operate either with a standard order- 
wire set or with a combination set composed of an operator's breast transmitter 
and the high-quality head-set used for monitoring. Direct monitoring is also pro- 
vided to a high-impedance head-set for use at the microphone boom. Provision 
has been made so that the mixer may talk to the boom man during a "take." 
The external appearance of the pick-up unit is shown in Fig. 3. 

Main Amplifier. The main amplifier is contained in a case measuring 18 X 10 
X 7 inches, and weighing 28 pounds. It consists of a two-stage amplifier having 
a double push-pull output stage. Tubes of a different type have been used in the 
two output stages, and, due to their characteristics, it is possible to supply the grids 
from a common input transformer, and have both sets of tubes reach the overload 
point at the same input voltage but at different outputs. One output stage sup- 
plies the light-valve, and the other the monitoring system. Both the low- and 
the high-pass filters are located in this unit, and may be cut into or out of the cir- 
cuit by means of switches. The gain of the amplifier is approximately 55 db., 
with a characteristic essentially uniform from 40 to 10,000 cps. The harmonic 
output of the amplifier at 1000 cps. is approximately 1 per cent in total harmonics 
at an output of +16 db. referred to the 0.006-watt level. 

Power Supply Unit. The power supply unit is contained in a case measuring 
18 X 14 X 10 inches, and weighing 137 pounds. Two of these units per channel 
are required, one supplying power to the pick-up unit and main amplifier, and the 
other to the recorder lamp and the noise-reduction and recorder control unit. 
Power consumption is moderate, being approximately 375 watts for each unit. 
For locations where extreme variations in voltage occur, a compact voltage regu- 
lator may be supplied. The appearance of the power unit is shown in Fig. 6. 

Gain Set and Oscillator. The portable gain set and oscillator is contained in a 
case measuring 18 X 13 X 10 inches, and weighing 59 pounds. It combines in 
one unit an oscillator and gain set which may be used as an adjunct to the record- 
ing channel, or may find considerable use in general studio maintenance work. 
The oscillator is variable over a frequency range from 40 to 14,000 cps. while the 
gain set will measure gains varying from to 120 db., or losses from to 10 db. 
Small plug-in impedance-matching units are provided for converting the gain set 
to operate with impedances other than 500 ohms. The equipment may be oper- 
ated from an a-c. power supply unit or from batteries. 

The oscillator is of the beat-frequency type, and the frequency dial has been 



196 



NEW MOTION PICTURE APPARATUS [J. S. M. p. E, 





FIG. 6. (Upper) Power supply unit. 
FIG. 7. (Lower) Portable gain set. 



SPEECH 
EQUALIZER 
MARGIN CONTROL 




COMBINED SPEECH AMPLIFIER 
AND DIODE RECTIFIER 




LOW PASS FILTER 
CONTROLS OPERATING 
SPEED 































OUTPUT 
TO L.V. 


20 KC 
OSCILLATOR 


CLASS C 
MODULATOR 


POWER AMPLIFIER 
I.2.0R3 VACUUM TUBES 


COPPER OXIDE, 
RECTIFIER 


RIPPLE 
FILTER 



FIG. 8. Schematic arrangement of noise-reduction system. 



Feb., 1937] NEW MOTION PICTURE APPARATUS 197 

calibrated in two ranges, one varying from 40 to 8500 cps., and the other from 
8000 to 14,000 cps. From 40 to 8500 cps. the output of the oscillator is uniform. 
The maximum output is +7 db. referred to the 0.006-watt level. The harmonic 
output of the oscillator is practically independent of the oscillator output, and con- 
tains approximately 1 per cent in second harmonic and 1 per cent in third har- 
monic for any frequency above 100 cps. Below 100 cps. the harmonic content is 
slightly greater. 

The oscillator is calibrated by one of two methods: The first combines a portion 
of the 50- or 60-cycle heater voltage with the output of the oscillator, producing a 
visible beat in the power level indicator of the gain set. The second makes use of a 
1000-cycle tuning fork to produce an audible beat with the oscillator output sup- 
plied to a receiver. The latter method may be used whether the oscillator is a-c. 
or d-c. operated. Frequency drift due to temperature changes will be found to 
be quite small after a short initial warming period. The frequency is also quite 
stable with respect to either A or B voltage variation. 

The gain set consists of a measuring and reference circuit connected to the out- 
put of the oscillator, and a receiving circuit. The measuring circuit contains 
various pads on keys, a variable attenuator, and a jack into which impedance- 
matching pads may be plugged. The reference circuit normally contains no at- 
tenuation, but for loss measurements the variable attenuator in the measuring 
circuit may be transferred to it by means of a key. The receiving circuit is en- 
tered through a jack into which impedance-matching units may be plugged. 
Preceding the power level indicator is another variable attenuator. The sensitiv- 
ity of the power level indicator may, therefore, be reduced according to the at- 
tenuation used. The power level indicator is of the high-speed type and is built 
out to 500 ohms, and has a sensitivity of 5 db. referred to the 0.006-watt level at 
mid-scale. The input-output key is arranged so that the circuit not connected to 
the power level indicator will be terminated in 500 ohms. A number of impe- 
dance-matching units are supplied to convert either the sending or receiving 
circuits to impedances other than 500 ohms, or from a balanced-H to a T type of 
circuit. These pads are stored inside the case. Fig. 7 shows the appearance of 
the oscillator gain set. 

Noise-Reduction. The general principles underlying noise-reduction have been 
previously described. 6 Briefly, a varying biasing current is generated which 
causes the mean spacing of the light-valve to open or close in proportion to the 
amplitude of the signal wave. This results in a relatively dark print during low- 
volume intervals, with consequent reduction in noise. 

The equipment described herein employs a modulated high-frequency carrier 
current. The advantages of this method are that: (1) the size of the units can be 
reduced, (2) the power efficiency is high, (3} circuit characteristics having opti- 
mal time-factors and filtering properties are readily obtained. 

The carrier-frequency method for noise-reduction has previously been used in 
a portable recording system. 1 The equipment herein described embodies further 
refinements of apparatus. The chassis of the unit slips into a duralumin case 
approximately 18 inches long, 10 inches wide, and 8Va inches high. The weight 
is about thirty pounds. 

The noise-reduction unit supplies 300 milliamperes or less to a 1-ohm light - 
valve circuit. The addition of one or two vacuum tubes in the output stage, pro- 



198 NEW MOTION PICTURE APPARATUS [J. S. M. p. E. j 

vision for which is made, increases the output current to as much as 900 milliam- 
peres if required. 

A 1000-cycle input signal of approximately 6 db. relative to 6 milliwatts 
cancels whatever value of biasing current may be set. Thus the biasing current 
may be changed without readjusting the margin control. The latter is a poten- 
tiometer having 0.5-db. steps associated with the input to the amplifier portion of 
the unit. 

The rate of change of biasing current when a signal is applied is such that the 
bias is reduced to 90 per cent of its initial value in 18 to 20 milliseconds. If sig- 
nals of amplitude greater than just sufficient to cancel are applied, the time is 
shorter. On removal of the normal input signal, the biasing current is restored 
to 90 per cent of its original value in 18 to 26 milliseconds. The filtering of un- 
wanted frequencies is such that when the noise-reduction is adjusted to 10 db., 
with 6-db. margin and half -cancelling biasing current, the noise components are 
38 db. below the signal at 50 cps. and 43 db. below at 100 cps. Extensive inves- 
tigation has been made of the relation of the rate of change of biasing current to 
the filtering, and the results attained with this equipment are believed to be near 
the optimal for single-track recording. 

The circuit arrangement of the new noise-reduction unit is shown in Fig. 8. 
An input signal from the main recording amplifier is first equalized in accordance 
with the frequency characteristic of the light-valve, then amplified, and subse- 
quently rectified by a duodiode-triode vacuum tube. The audible-frequency 
components are filtered to the maximum that will permit an optimal rate of 
change of biasing current. The output voltage is of an amplitude proportional 
to the envelope of the recorded signal, and is impressed upon a class C modulator 
stage to which is also connected the 20-kilocycle output of a vacuum tube oscilla- 
tor. The circuit characteristics are such that the amplitude of the 20 kilocycles 
transmitted through the modulator stage is in inverse proportion to the voltage 
output of the rectifier filter. The modulator output is coupled to a power-ampli- 
fier stage utilizing a pentode and then to a copper-oxide rectifier. The high- 
frequency ripple is filtered out, and the final current transmitted to the light- 
valve simplex circuit. 

Film Recorder. The fundamental requirement of any type of film recorder 
should be to propel film past the modulating light-beam at constant speed. For 
a studio type of recorder, where weight is of no importance, the problem of speed 
constancy is materially aided by designing the moving parts, such as shafts, 
rollers, flywheels and driving motor, considerably oversize, and thus allowing rea- 
sonable tolerances in manufacture. For a portable recorder such a procedure is 
obviously undesirable. It is necessary, therefore, to analyze the entire structure 
carefully, and drastically eliminate weight wherever possible. This, in turn, 
places a much more severe requirement upon the mechanical precision through- 
out. It also precludes the incorporation of unessential automatic devices. 

The film recorder used with this channel is shown in Fig. 9. The film path is 
quite short and simple. A sprocket pulls the film from the magazine at the left, 
whence it is passed on, with a free loop intervening, to an impedance roller of con- 
siderable mass. The recording is done on this roller, which is driven only by the 
film as the latter is pulled by the filtered pulling sprocket, located approximately 
n the center of the film compartment. As the film passes from the impedance 



Feb., 1937] NEW MOTION PICTURE APPARATUS 199 

roller to the pulling sprocket, it comes into light contact with a fixed shoe partially 
surrounding the pulling sprocket. The friction between the film and this shoe is 
instrumental in eliminating sprocket-hole disturbances. 

The film is guided from the filtered sprocket by two idler rollers, one fixed and 
one movable, which act as strippers, to minimize sprocket-hole disturbances. The 
film passes from these stripper rollers to the hold-back side of the upper sprocket, 
through another free loop. From the hold-back sprocket the film passes over 
rollers into the take-up side of the film magazine. 

It should be particularly noted that the filtered sprocket is isolated from the 
pull-down hold-back sprocket by means of two free film loops. The only impe- 
dance offered to the filtered sprocket is therefore that of the impedance roller. As 
this roller has considerable inertia, any high-frequency sprocket-hole fluctuations 
are effectively attenuated. 

Lower-frequency disturbances due to mass unbalance and inaccuracies of gear 
teeth are most effectually eliminated by means of a mechanical filter. In this 
portable recorder, the mechanical filter is similar to that used in the standard 
Western Electric studio recorder. It consists of a driven gear coupled to the 
sprocket shaft by means of springs. Upon the sprocket shaft is mounted a heavy 
flywheel. The elasticity of the springs in combination with the mass of the fly- 
wheel, plus the proper amount of frictional damping, constitutes a filter providing 
extremely constant rotational speed of the filtered pulling sprocket. 

The sundry gearing, along with the spring coupling of the mechanical filter, are 
placed at the rear of the recorder housing in an oil-tight compartment. To elimi- 
nate flexible couplings, and also to conserve space, the driving motor projects into 
the gear compartment. Two worms (actually one unit) are mounted directly 
upon the tapered end of the motor shaft. The larger of the two worms drives 
downward to the worm gear coupled to the filtered sprocket shaft assembly, 
whereas the smaller worm drives upward to the pull-down hold-back sprocket 
assembly. The end of the motor shaft also couples to the footage counter drive. 

The spring coupling and friction damping assembly are surrounded by a light 
metal cover. The oil level is such that the metal cover dips well into the lubri- 
cant ; and as the cover rotates it carries considerable oil upward, by passing it to 
the driving worms mounted upon the motor shaft. Thus an adequate supply of 
oil is always assured for the gears. All the bearings, except those used in the 
footage counter assembly, are of the precision oil-sealed ball-bearing type. When 
the recorder is in the normal position the level of oil is always well below any of 
the bearings, so that no hydrostatic pressure is exerted against any bearing con- 
necting to the outside of the oil reservoir. This insures the film compartment 
against leakage of oil. For positions other than normal, the oil chamber is so de- 
signed that only a very small amount of oil can be in contact with any bearing, 
and thus the leakage through the bearing, if any, is insignificant. 

To assist in assembling or disassembling the recorder motor and gearing, a cir- 
cular clear celluloid window has been provided at the top of the recorder case which 
is normally covered by a quickly removable cover plate. This window is also a 
means of inspecting the oil level. Since all lubrication is provided automatically, 
maintenance is considerably reduced. 

By removing the cover from the right end of the recorder, Fig. 10, the modulator 
unit, driving motor, transformers, resistances, and jacks are exposed to view. 



200 



NEW MOTION PICTURE APPARATUS [J. S. M. P. E. 




FIG. 9. ( Upper} Film recorder, magazine film compart- 
ment and modulator door open. 

FIG. 10. (Center) Film recorder, right front end, with 
cover plate removed. 

FIG. 11. (Lower) Recorder control unit. 






Feb., 1937] NEW MOTION PICTURE APPARATUS 201 

The modulator unit uses the permanent-magnet type of light-valve. The 4-am- 
pere exciter lamp is mounted in a quickly adjustable lamp mount. 

The motor shown in Fig. 10 is of the 12-volt d-c. interlocking type. However, a 
110-volt d-c. interlock motor, as well as regular synchronous and standard Western 
Electric interlock motors, are available for this recorder. 

Photoelectric cell monitoring is effected by deflecting some of the stray modu- 
lated light, by means of a curved mirror, downward to a small photocell 
mounted in the base of the machine. This photocell is coupled directly to a two- 
stage amplifier located in the left end of the recorder. The output of this am- 
plifier, as well as the other electrical circuits for the recorder, except that of the 
motor, are carried, by means of short cables, down to the recorder control unit. 

The recorder control unit houses all the necessary switches and meters to oper- 
ate or control the recorder except those connected with the driving motor. As 
any one of several types of motor may be used, it would unnecessarily complicate 
the recorder control unit to supply adequate control facilities for all of them. 
Consequently, the motor control is located in a small switch box, which is sup- 
ported from the side of the recorder during operation. 

The recorder in Fig. 9 is shown with a Mitchell magazine. By means of an 
adapter plate, which can be permanently mounted upon the recorder if desired, 
standard Bell & Howell magazines can be used. The take-up drive is through 
either a spring or leather belt. Ordinarily either type of belt would slip suffi- 
ciently on the magazine pulley to avoid the necessity of a separate friction take-up 
clutch. As a safety precaution, however, a clutch has been provided, adjustable 
by means of three screw-pins mounted in the face of the pull-down and hold-back 
sprocket. 

In case the take-up fails for any reason, a film buckle release switch has been 
provided for disconnecting the motor from the power supply following the forma- 
tion of a small loop of film next to the take-up sprocket. 

The overall dimensions of the recorder are 17 6 /ie X 9"Ae X ll*/i inches. The 
weight is approximately 100 pounds, although by removing the filter flywheel 
about 15 pounds can be eliminated. Handles are provided so that the recorder 
can readily be moved about upon a set. 

The recorder is finished in a gray lacquer that is quite resistant to rough usage 
and can easily be cleaned by wiping whenever necessary. 

Speed Regulation. For accurate speed control a portable speed-regulating unit 
has been provided. When using either the synchronous or the standard Western 
Electric interlock motor, no additional speed-regulating equipment is necessary. 
For recording work not requiring very exact speed regulation, a 12-volt d-c. driv- 
ing motor may be used, the speed being manually adjusted by means of a speed 
indicator. The change from this set speed during a take is very slight as the power 
is supplied by storage batteries. 

Recorder Control Unit. The recorder control unit is contained in a case 18 X 10 
X 11 inches in size, weighing 42 pounds. It contains the various switching facili- 
ties required to control the recorder lamp, the noise-reduction unit, and the light- 
valve. Special features contributing to ease of channel operation include an os- 
cillator for checking light-valve overload and tuning, and a bridge circuit for de- 
termining valve overload. The appearance of the unit is shown in Fig. 11. 

The units that have been described form a compact, mobile sound-recording 



202 NEW MOTION PICTURE APPARATUS [J. S. M. P. E. 

system, capable of producing high-quality sound records. Operation of the chan- 
nel has been simplified and made flexible by the use of a number of new features. 
Maintenance is minimized by the easy accessibility of the parts and by the fact 
that any unit may be quickly removed from the channel and another substituted 
for it. 

REFERENCES 

1 DAILY, C. R.: "A New Western Electric Double-Film Portable Sound Re- 
cording System," 7. Soc. Mot. Pict. Eng., XX (Feb., 1933), No. 2, p. 128. 

2 PEARSON, G. L.: "Quiet Amplifier Tubes," Bell Laboratories Record (Oct., 
1935), No. 2, p. 56. 

3 GANZ, A. G., AND LAIRD, A. G.: "Improvements in Communication Trans- 
formers," Bell Syst. Tech. J. (Jan., 1936), No. 1, p. 136. 

4 MARSHALL, R. N., AND ROMANOW, F. F.: "A Non- Directional Microphone," 
Bell Syst. Tech. J. (July, 1936), No. 3, p. 405. 

6 Progress in the Motion Picture Industry : Report of the Progress Committee, 
J. Soc. Mot. Pict. Eng., XXVII (July, 1936), No. 1, p. 22. 

6 SILENT, H. C., AND FRAYNE, J. G.: "Western Electric Noiseless Recording," 
/. Soc. Mot. Pict. Eng., XVIII (May, 1932), No. 5, p. 551. 



A NEW HIGH-QUALITY FILM REPRODUCER* 
J. C. DAVIDSON** 



Talking motion pictures became a commercially successful fact in August, 
1926, and many of the original Western Electric reproducing equipments in- 
stalled soon after that date are still in daily use. These machines were the 
result of a careful analysis of the requirements of the industry as they were then 
understood. 

During the past two years, surveys have been made, concurrently with develop- 
mental work, with a view to the production of a reproducer that will meet not 
only today's requirements as we see them, but will accommodate future develop- 
ments in recording technic for several years to come. 

Requirements. In order to give to the theater owner full measure of improve- 
ment, together with the years of service he is entitled to expect, the film propelling 
mechanism should be capable of moving the film past a scanning point with a 
degree of constancy of speed substantially equal to that attained in the more 
costly studio recording equipment. This accomplished, the equipment should 
be able to reproduce, without distortion, a far greater volume range than is at 
present attainable. Machine noise should be reduced to a degree that will 

* Received October 10, 1936; presented at the Fall, 1936, Meeting at Roches- 
ter, N. Y. 

** Electrical Research Products, Inc., New York, N. Y. 



Feb., 1937] 



NEW MOTION PICTURE APPARATUS 



203 



permit deriving full advantage from the quiet passages, and, in addition, the 
reproducer should be ruggedly constructed, easy to operate, and simple to main- 
tain. Parts in which unavoidable wear occurs should be readily replaceable. 
As a whole, the machine should be a complete, compact, symmetrical unit, and 
should not present the appearance of an appendage to its associated projector 
head. 

Mirrophonic Reproducer Set. The Mirrophonic heavy duty reproducer set 
has been designed to fulfill the above-described requirements. It consists of 
three units which interchangeably fit together to form a single symmetrical unit. 

The main film compartment consists of an accurately machined casting con- 
taining the film-propelling mechanism, gears, the kinetic scanner, and a sub- 




FIG. 1. View of film and photocell compartments. 



compartment in which the exciter lamp is located. The photocell compartment 
is attached to the rear end of the main casting (and is interchangeable). It 
contains the photocell, the scanning slit, and coupling transformer. Fig. 1 is a 
view of the film and photocell compartments. The motor assembly (Fig. 2) is 
attached to the forward end of the main casting as a unit, and is interchangeable, 
self-aligning, and can be replaced within a few minutes. 

The film path is from the hold-back sprocket in the projector head, around the 
drum of the kinetic scanner, over the drive and hold-back sprockets, and thence 
to the lower magazine. The mechanical drive consists of a worm directly coupled 
to the motor, driving a gear at 360 rpm. The gear is on the cross-shaft that sup- 
ports the main drive sprocket as well as the necessary gears for driving the pro- 
jector head and the lower magazine take-up. The cross-shaft, as well as the 
worm shaft, is supported in sealed ball bearings, and their accuracy of alignment 
is held to very close tolerances. The cross-shaft is coupled to the hold-back 



204 



NEW MOTION PICTURE APPARATUS [J. S. M. p. E. 



sprocket shaft by means of a set of steel and fiber gears ; and in order to minimize 
noise, the lower magazine take-up is driven by a silent chain. 

The lubrication is fully automatic. All shafts but the one bearing the hold- 
back sprocket, and including the motor shaft, rotate in sealed ball bearings. 
The worm gear operates in oil in a sealed chamber in the main casting, and the 
hold-back sprocket shaft is rifle-drilled to permit lubrication from this chamber. 
The kinetic scanner (Fig. 3) is a completely sealed unit, consisting of a hardened 
nitralloy scanning drum, ground to a concentricity of better than 0.0001 inch, 
rotating on a shaft running in sealed ball bearings. It is a complete unit, mounted 
in the main frame casting. A two-element film speed governor, mounted upon 
the rear end of the shaft, insures uniform speed of film propulsion. 

The film is held in contact with the scanning drum by a pressure pad roller 

which serves also to maintain scanning 
alignment. This assembly consists of a 
shaft, mounted on ball bearings, upon 
which is a felt pressure roller built up of 
a series of felt rings cemented together, 
thus assuring a uniform hardness of sur- 
face capable of maintaining its original con- 
centricity. The felt roller is mounted upon 
a steel sleeve and is easily replaceable as 
a unit. 

The optical system (Fig. 4) is the projec- 
tion type of scanner. An exciter lamp is 
mounted in a pre-focused lamp bracket 
having the usual adjustments, which is 
mounted upon a damped chassis merely by 
being pushed upon two locating studs and 
locked into place. The light is focused 
upon the film by a condenser and prism 
combination mounted in a slot in the 
main frame casting which is adjustable 
along the optical axis for optimum setting. 
The optical axis is fixed throughout by 
the exceedingly close manufacturing tolerances employed. 

The objective lens is a standard microscope objective mounted upon a pre- 
cision sleeve that provides movement along the optical axis for obtaining a sharp 
focus. This objective projects an image of the sound-track, magnified ten times, 
upon a scanning slit in the photocell compartment. Provision has been made for 
masking the width of the track scanned at the focal plane, thereby insuring 
against the decrease of illumination at the edges of the beam that occurs when 
masking is attempted at points other than on the focal plane. The height of 
the slit is adjusted at the factory to be an optimum for the frequency range to 
be scanned. The azimuth of the scanning slit is readily adjustable. 

The principal immediate advantage of projection scanning is its flexibility 
and simplicity of adjustment. It bids fair to take the scanning system out of the 
hands of the optical laboratory and place it with the projectionist, who will find 
the adjustments as simple as those he has always been accustomed to make in 




FIG. 2. The motor assembly. 



Feb., 1937] 



NEW MOTION PICTURE APPARATUS 



205 



projecting pictures upon the screen. The magnified image of even an 8000-cycle 
sound-track is so large that focus and azimuth can readily be adjusted by visual 
inspection, and will not be more than a decibel or so from the optimum obtained 
by the more laborious method employing the film loop and volume indicator. 
A septum is provided at the center of 
the scanning slit for the reproduc- 
tion of push-pull recordings. It is 
mechanically connected to the switch 
controlling the electrical circuits for 
reproducing either standard or push- 
pull sound-track. 

The electrical circuit is in the photo- 
cell compartment, and consists of the 

i photocell (which may be of the usual 
single-element type for standard track, 
or double element type for either push- 
pull or standard track) and a carefully 
balanced and shielded transformer 
having an impedance ratio of some- 

i what more than a third of a megohm 
to 200 ohms. The electrical balance 
between the elements of the cell and 
the coils of the transformer is such 

i that no provision is required for 

i equalizing the output from the two 

| halves of a push-pull sound-track, so F IG> 3. The kinetic scanner. 

jthat operation is simplified merely to 

I throwing a switch for the type of sound-track to be scanned. 

From the installation and operating standpoints, the question of simplicity 
has been given serious consideration. With the selection of the correct pedestal 





FIG. 4. The optical system. 



iarm, the reproducer set may be readily mounted upon any of the pedestals manu- 
factured today in the U. S. A. Adapters are provided for the various projector 
! heads in current use. The projector head is fastened to the adapter, which slides 
into a groove in the top of the main frame casting, providing a simple means of 
correctly meshing the reproducer and projector gears. It is possible, therefore, 
i to remove the projector head, together with the adapter, by removing four readily 



206 NEW MOTION PICTURE APPARATUS [J. S. M. P. E. 

accessible bolts, and to mount the projector head again without in any way chang- 
ing the focus or alignment of the picture upon the screen. 

Performance The performance of this reproducer set fulfills all requirements 
indicated by present-day standard of reproduction of sound recorded on film, and 
also anticipates any future developments that at present can be foreseen. The 
flutter content of the average machine, measured in production, is about 0.1 per 
cent; and with special adjustment the machine is capable of bettering this per- 
formance. The frequency characteristic conforms to the theoretical response for 
a scanning beam of the height employed. The introduction of calculated damp- 
ing materials insures that the machine introduces no noise during the quiet pass- 
ages of sound-track. 

The reproducer set has been designed to have the appearance of a complete 
symmetrical machine. The inside is finished in white, to promote cleanliness and 
provide better visibility for threading the film. The materials used, the care in 
manufacture, and the finishes applied are all of the best, and it is confidently 
felt that the new reproducer set will give even longer service than its prede- 
cessors, many of the earliest of which are, as already mentioned, still in daily 



RECENT DEVELOPMENTS IN HIGH-INTENSITY ARC SPOTLAMPS 
FOR MOTION PICTURE PRODUCTION* 



E. C. RICHARDSON** 






The high-intensity carbon arc affords certain advantages as a light-source for 
photography that are not possessed by other illuminants. Within the restricted 
area of its positive carbon is concentrated an intrinsic brilliancy greater than that 
afforded by any other artificial light-source. Fortunately, the distribution of 
radiant energy throughout the spectrum of a high-intensity carbon arc coordi- 
nates well with the spectral sensitivity of photographic emulsions and the trans- 
mission factors of camera lenses. 

For the purpose of more effectively utilizing high-intensity arc sources in mo- 
tion picture photography, two lamps have been recently developed. The M-R 
Type 90 lamp (Fig. 1) operates at 120 amperes. The M-R Type 170 lamp (Fig. 2) 
has a capacity of 150 amperes. The designs of these two lamps, which are in 
general quite similar, embody many new factors that greatly enhance their utility 
and add to the convenience of operating them. Fig. 3 shows the mechanism of the 
Type 90 high-intensity arc lamp, in which the following vital improvements have 
been incorporated: (1) increased rotational speed of positive carbon; (2) continu- 
ous non-intermittent feeding of both positive and negative electrodes; (3) rapid- 
action positive and negative manual adjustments. 

* Received September 8, 1936; presented at the Fall, 1936, Meeting at Roches- 
ter, N. Y. 

** Mole-Richardson, Inc., Hollywood, Calif. 



eb., 1937] 



NEW MOTION PICTURE APPARATUS 



207 



When these new high-intensity arc lamps were being designed it was known that 
hey would find extensive use in color motion picture photography. The require- 
ments of illuminants for color cinematography are very rigid as to uniformity of 
pectral distribution and intensity. Considerable experimenting was done with 




FIG. 1. (Left) MR Type 90 high-intensity arc spotlamp (rear view). 
FIG. 2. (Right) MR Type 170 high-intensity arc spotlamp (front view). 



arious rotational speeds of positive electrode. It was found that at approxi- 
mately 14 rpm. an optimal condition was established and that the crater was very 
jymmetrical. It was noted that at that speed crater rims that had been chipped, 
iither by careless striking or other causes, were quickly restored to symmetry. 
| The maintenance of a properly shaped, symmetrical crater is one of the most 
jnportant requirements for stability of the arc. In the arcs under discussion, 
le angle between the positive and the negative electrodes is 127 degrees. It has 



208 



NEW MOTION PICTURE APPARATUS [J. S. M. P. E. 



been found that at that angle practically any carbon suitable for use under high- 
intensity conditions performs reasonably well. The 127-degree angle results in 
better performance than the much natter angle generally used in the design of 
searchlights and Sun-arc lamp mechanisms. The objection to this angle is that 
in striking the negative to the positive, the contact between the two electrodes 
comes at the rim of the positive crater, which tends to cause chipping. That was 
one of the reasons why the lamp was designed to be struck manually rather than 
automatically. The other reason was that automatic striking would have in- 
creased the cost of the mechanism to a very considerable extent. 




FIG. 3. Mechanism of MR Type 90 high-intensity arc spotlamp. 

The feeding arrangements of these new high-intensity arc mechanisms diffe 
from other designs in combining continuous, non-intermittent feeding of the posi 
tive and negative electrodes with rapid positive electrode rotation. So far as th 
writer is aware, all high-intensity arc mechanisms used in motion picture photog 
raphy, having rotational speeds sufficiently rapid to maintain good crater condi 
tions, have had feeding arrangements of the "stop and go" type. Only by unii 
form, non-intermittent feeding and positive electrode rotation faster than 10 rpirj 
can the constant stability of performance be attained in high-intensity arcs usei 
for illumination in color photography. The mechanics of the feeding arrange! 
ments in these lamps permitted a rapid hand control: one revolution of th! 
negative hand-feeding crank advances the negative electrode 0.1 inch. A simila 
movement of the positive hand control moves the positive electrode 0.08 inch. 



Feb., 1937] 



NEW MOTION PICTURE APPARATUS 



209 



Quietness of operation is, of course, essential in all motion picture equipment de- 
signed for use in conjunction with sound-recording apparatus. The motors of 
these new high-intensity arc lamps have grease-packed reduction gears, and no 
shafts or parts other than the motor reduction gearing have rotational speeds 
greater than 46 rpm. All shafts and rotating members, other than those rotating 
at the slowest speeds, are mounted upon either oil-less or ball-bearings. Again 
and again, in recent productions, these lamps have been operated satisfactorily 
within six or eight feet of the microphones. Provision has been made for com- 



400 



1600 



1800 



800 



400 



-! 



8 

i A. 10 degrees 

fe B 18 

** C 84 

D 52 

I 44 



38 




Type 90 
High-Intensity Arc 

115 Volta 
110 uaperes 
60 Aro Volts 



12 8 4 4 8 

Divergence in degrees 



12 16 20 24 



FIG. 4. Typical distribution curves of Type 90 high- 
intensity arc. 

pletely stopping the driving motor in the few situations that arise when the mi- 
crophone has to be placed within six feet of the lamp. 

Aside from improvements in the mechanism, probably the most outstanding 
Pther improvement in these new equipments is the application of "Morinc" flat 
(corrugated lenses as a means of collecting and projecting the light of the arc. The 
[carbons used in the Type 90 are a 13.6-mm. positive and a 3 /s X 9-inch copper- 
poated negative. The Type 170 uses a 16-mm. positive carbon and a Yie-inch 
copper-coated negative carbon. It is characteristic of these high-intensity arc 



210 



NEW MOTION PICTURE APPARATUS IJ. S. M. P. E. 



combinations that the most effective portion of their radiation falls within an 
angle of 60 degrees each side of the axis, and principally within a total angle of 80 
degrees. Within 80 degrees the intensity is not less than 10 to 17 per cent of the 
intensity at the axis. Heretofore, in order to project the light from a high-in- 
tensity motion picture arc spotlamp effectively, the principal optical means em- 



laoo 



800 




18 8 4 48 

Divergence in degrees 



FIG. 5. Typical distribution curves of Type 170 high 
intensity arc. 



ployed have been parabolic reflectors and spherical lenses. The high tempera- 
ture developed directly in front of the positive crater in arcs operating at currents 
greater than 100 amperes is such that lenses, for practical use, have been limited 
in diameter to 8 inches, and to focal lengths such that the arc would not be closer 
to the lens than 6 inches when the desired flooding angle was attained. Experi- 
ence has demonstrated that if lenses greater than 8 inches in diameter were used, 
or if the arc were brought closer than 6 inches, the glass elements were subjected 
to pitting by the copper coating of the negative electrode and the hazard of break- 



Feb., 1937] NEW MOTION PICTURE APPARATUS 211 

age was greatly increased, regardless of the efforts of lens manufacturers to make 
use of the advantages of heat-resisting glass having low coefficients of expansion. 

The 24- and 36-inch parabolic mirrors, which have given the best performance, 
have both been limited to the 18 8 /4-inch focus. The mechanical limitation has 
been imposed by the fact that there must be room for interposing a negative carbon 
of suitable length between the crater and the mirror surface when the lamp is ad- 
justed for flooding. This prevents the positive crater from being placed nearer 
than ll s /4 inches from the center of the mirror surface. Even though the lamp 
operators on the stages exercise the utmost care to protect their sun arcs from 
wind and sudden changes of temperature, there has always been a great deal of 
dissatisfaction regarding the breakage of parabolic mirrors in studio use. The 
"Morinc" lens, designed for these new high-intensity arc lamps affords ideal dis- 
tribution of illumination for photographic purposes. Figs. 4 and 5 illustrate the 
distribution attained at various angles of divergence when the lamps are used for 
flooding, and show the wide range of divergence attainable. Variations from spot 
beams of 8 degrees to floods of 44 degrees are produced having smooth fields of 
photographically useful light. The edges of the various beams vignette and 
make it practicable to overlap fields of illumination, as is often necessary in motion 
picture set lighting, without creating high-intensity areas in the overlaps or dis- 
tinct markings defining the circumference of the field. 

In both lamps every effort has been made to facilitate operating them in their 
many applications in the studios. Each has a resistance grid, so designed that it 
may be removed from the pedestal, permitting the grid and lamp head to be taken 
as a complete operating unit to the overhead cat-walks and parallels. 

The Type 90 lamp weighs 224 pounds and is tending to replace the 24-inch Sun 
arc, which as a rule weighs more than twice as much. The Type 170 lamp, which 
weighs 311 pounds, is, in most cases, replacing the 36-inch Sun arc, many of which 
weigh considerably more than 600 pounds each. 

In recognition of the value of reflector arcs of the type generally classed as Sun 
arcs, it should be noted that for divergences of less than 15 degrees they show 
definite superiority, and it is not anticipated that projectors of the lens type will 
displace them for uses requiring such narrow beams. 

Lamps of the types described have carried the major burden in all artificial 
lighting used in the Technicolor productions Trail of the Lonesome Pine, Dancing 
Pirates, Garden of Allah, Ramona, and God's Country and the Woman. Very con- 
siderable numbers of lamps are now in use at the Warner Brothers, Metro-Gold- 
wyn -Mayer, Paramount, United Artists, and Columbia studios in Hollywood, and 
at London Films Denham Studio in England for both black-and-white and color 
photography. 

DISCUSSION 

MR. PALMER: Is the motor connected directly across the line, so that the speed 
does not vary with arc voltage? 

MR. RICHARDSON: It is a shunt motor, connected across the arc. We experi- 
j mented with placing the fields across, or ahead of, the resistance, but adopted the 
present arrangement because we have to separate the rheostat from the lamp head 
on account of the heat. The motor changes speed somewhat with the slight volt- 



212 NEW MOTION PICTURE APPARATUS [J. S. M. p. E. 

age fluctuations resulting from changes of arc length, so we have some degree of 
regulation. 

The mechanism will work the positive carbon down to about 3 5 / 8 inches, when 
consumed to the limit. There have been many attempts to devise ways of saving 
carbons; but using such heavy currents as we do in the high-intensity lamps, per- 
fect conductors are required to get the current to the tip of the positive carbon. 

MR. MISENER: Are the motors sealed? 

MR. RICHARDSON : They are entirely enclosed, and require very little servicing. 
In fact, I do not believe that, of 250 lamps we have hi operation, more than 
10 have ever required servicing. Eventually brushes will have to be replaced; 
but if the motor is used, say, two hours a day, the brush life should be quite 
long. 

MR. CRABTREE : We read in the newspapers about the high temperatures ex- 
isting in the Hollywood studios. What is being done in the way of air condition- 
ing? 

MR. RICHARDSON: That question has been asked a number of times. I have 
not heard of the terrific heat, except that the temperature does get rather high at 
the studios in the San Fernando Valley. The outdoor summer temperature there 
is occasionally 100 or over. As far as I know the stages are not refrigerated, al- 
though many of them pass the air through water sprays, which have some cooling 
effect. 

MR. TASKER : Many stages are now equipped with blowers for changing the air 
during takes. There are three conditions under which the heat becomes uncom- 
fortable: when one is working in a small closed room, particularly for photo- 
graphic reasons; when the actors are wearing arctic or whiter costumes on the 
stages; and when working with color. 

MR. RICHARDSON: A large cold-storage company in Hollywood had an enor- 
mous ice storage building that had become obsolete due to the activities of the 
electric refrigerator manufacturers. They conceived the idea of turning it into a 
cold stage, and installed refrigerating equipment. When you see Columbia's 
Lost Horizon you will see the actors walking about with vapors coming from their 
nostrils a very convincing arctic scene, because the temperature may be as low 
as 20 degrees, even on a summer's day. 



THE SCHWARZKOPF METHOD OF IDENTIFYING CRIMINALS* 

J. FRANK, JR.** 

During the past two years experiments have been conducted that have resulted 
in a recognized contribution to the science of sight identification. This paper de- 
scribes the background indicating the need for such a method, some of the ex- 

* Received October 9, 1936; presented at the Fall, 1936, Meeting at Roch- 
ester, N. Y. 

** International Projector Corp., New York, N. Y. 



Feb., 1937] NEW MOTION PICTURE APPARATUS 213 

periments conducted, the equipment and technic finally agreed upon, and some of 
the results to date. 

There are two systems of identifying human beings : sight and positive. The 
law enforcement officers must use both types in apprehending criminals. The 
positive system is applicable only when a criminal is in custody. It consists in 
taking measurements, recording tattoo marks, scars, and other features, and 
comparing the records with any previous records the individual may have. Fre- 
quently during the process of establishing the records for positive identification, 
full-face and profile still pictures are taken of the individual. 

Identification by sight is, on the other hand, applied to criminals who are at 
large and are wanted by the police. The field of development of sight identifica- 
tion has been sadly neglected. It is necessary for the law enforcement officer to 
identify a wanted criminal by sight first, and then take him into custody before 
positive identification can be effective. With improved modes of transporta- 
tion, improved methods of disguise, and increasingly larger crowds in cities, 
more highly scientific methods of sight identification are required. 

In the earlier days, law enforcement officers had only verbal descriptions of 
the wanted criminal to guide them. A few of the officers might have actually 
seen the criminal when he was previously in custody. Then came the printed 
description, with reproductions of photographs, the familiar "Wanted" circular, 
with its full-face and profile views. The "Wanted" circular, because it is in- 
expensive and can be widely and quickly distributed, is a useful tool. At the 
very best, however, the halftone reproductions on the printed circular are several 
steps removed from the original photograph, and represent the subject as he ap- 
peared during only a fleeting moment the fraction of a second during which the 
camera exposed the film and during which the criminal may have distorted his 
countenance somewhat. 

To compensate for the deficiencies of the still-picture likenesses, modern large 
city police departments use the "line-up." The advantages of the line-up are 
that the police see the criminals in action; they see his characteristic gestures, 
the way he stands and walks; and hear his voice. This is the most effective 
means of promoting sight identification. The disadvantages of the line-up are 
that only a handful of all interested law enforcement officers can be present at the 
line-up while the criminal is in custody. 

As the modern and most effective method known of familiarizing law enforce- 
ment officers with the identifying characteristics of wanted persons, the New Jer- 
sey State Police, working in collaboration with the RCA Manufacturing Co., 
the J. M. Wall Machine Co., and the International Projector Corp., have de- 
veloped a method of using talking motion pictures the Schwarzkopf method of 
identifying criminals. Experiments were made using 16-mm. sound motion 
picture equipment, which demonstrated the great value that such records would 
have in assisting in establishing identification by sight. Further development 
was carried on with 35-mm. equipment of both the single- and double-film type 
over a period of more than six months. 

Because the films are to be used for identification, the first requirement for 
police work is realism in the projected picture and reproduced sound. The 
pictures must have clear, sharp details, and the sound must reproduce in recog- 
nizable form the inflection, pitch, and other characteristics of the subject's voice. 



214 



NEW MOTION PICTURE APPARATUS [J. S. M. p. E. 



The equipment must be so simple that one person can operate it without con- 
siderable training, and must be so designed as to be easily transported and 
quickly set up at any place where electric current is available. 

The experiments resulted in the decision to use, for high-quality results, double- 
film recording equipment originally designed for location and industrial purposes. 
For certain otjier applications t<he 16-mm. equipment fulfilled the requirements. 
Further development of a single-film system employing a noise-reduction system 
will result in a satisfactory set-up at lower cost. 

Recording equipment, camera, and illuminating equipment were modified 
for this particular application. Ease of focusing and suitable acoustical condi- 
tions were attained by using a floor mat of canvas and gray cloth drapes forming 
a triangle open at the apex. The floor mat has fixed sockets for the legs of the 




FIG. 1. Complete set-up of equipment and stage. 



camera tripod, and the proper location for the subject is marked so that he will 
not get out of focus or out of the range of the camera lens. The side and back 
drapes provide a suitable photographic background, concentrate attention upon 
the subject, and eliminate echoes or reverberations that might be picked up by 
the microphones. The positions and the wattages of the necessary lamps are 
predetermined to provide the proper lighting. 

Setting up the equipment requires perhaps fifteen minutes. When all is ready 
and the subject is placed in the position marked upon the floor mat, with the 
microphone suspended immediately above and in front of him, the inquisitor 
takes his position at the other microphone and the operator standing at the 
camera presses the control button that he holds in his hand. At once the subject 
is brightly illuminated by the flood lamps. To allow the subject to regain his 
composure, and to give the operator time in which to glance at his lights and 
meters, the automatic controls have been so arranged that a lapse of three seconds 



>., 1937] 



NEW MOTION PICTURE APPARATUS 



215 



between the instant the lights go on and the instant the camera and re- 
are started. The inquisitor then puts his questions and directions to the 
ibject, according to a prearranged plan, so that the interview requires three 
mtes and uses approximately 300 feet of film. Three minutes has been found, 
considerable experimenting, to be sufficient time in which to describe the 
ibject from four different angles, using long, medium close-up, and close-up 
as well as to ask the customary questions and have the subject perform 
tain distinguishing actions. 
In most cases, portable reproducing equipment will fulfill all requirements, 




FIG. 2. Stage and equipment in use. 

with the advantage that showings may be made in various locations. In some 
police headquarters in the larger cities it may be desirable to have one of the per- 
manent installations such as is supplied to theaters. 

The possibilities of the Schwarzkopf method of identifying criminals indicate 
that it is an important step forward in police methods. It is the plan to locate 
recording apparatus in the large cities, and in the possession of State police au- 
thorities as required to meet the needs of the State. Each State and Federal 
prison should be likewise equipped. Sound motion pictures of criminals should 
be made when they are apprehended, when they arrive at prison if convicted, and 
immediately before their release, to insure getting accurate likenesses of the 
prisoners' voices and customary appearances. Habitual criminals and those sus- 
pected of more serious crimes will be photographed and recorded upon film as a 



216 NEW MOTION PICTURE APPARATUS [J. S. M. p. E. 

matter of routine, just as they are now fingerprinted. At regular intervals police 
officers will be shown sound motion pictures of all criminal persons known to be 
in the vicinity. 

The police departments of even the smaller towns are to have reproducing 
equipment, either for standard 35-mtn. or for 16-mm. film. Large cities should 
have 35-mm. reproducers in each precinct police station as well as at headquarters. 
Instead of having only a few headquarters' police officers and detectives see the 
criminals when they are picked up, every law enforcement officer, from the patrol- 
man on the beat to the highly trained detective, can see and hear the wanted 
persons, not after they are in custody, but when they are wanted; and not only 
in the headquarters of the police force that arrested them, but anywhere or every- 
where they are likely to be. They can see and hear the picture monthly, weekly, 
or daily if they desire. 

Undoubtedly, in these days of fast travel, when criminals roam from one end 
of the country to the other, some means of coordinating the sound-film activities 
of local police departments will be needed. Probably this may be achieved by 
a national library, which will maintain files of films just as fingerprint files are 
now maintained. Local police would record the criminal on film and have made 
whatever number of projection prints may be required. The original negative 
would then be sent to the national library, which would supply prints, either 
35-mm. or 16-mm., to the various police departments as needed. 

In the cases of the more notorious criminals, their pictures could be shown even 
in the theaters of the country, as part of the newsreels. Newsreel producers 
have already indicated their eagerness for such films. What chance of escaping 
detection would a "public enemy" have when every police officer and millions of 
citizens have seen him in action and have heard him talk? 

A demonstration was held on May 27, 1936, at Trenton, N. J., under the spon- 
sorship of Gov. Harold G. Hoffman before an audience of 500 persons representing 
the most important law enforcement agencies in the United States, England, and 
Canada. Four pictures of actual criminals in New Jersey were made for this 
demonstration. 

On August 25th, John A. Byers escaped from the Middlesex County workhouse. 
It happened that Byers was one of the four criminals whose pictures had been 
taken. Before it was possible to complete arrangements to use the picture to 
assist in capturing him an extraordinary thing happened. 

On September 5th, Byers telephoned to a friend from a pay-station in Asbury 
Park. County Detective Drosdick, a former state trooper, who had never ac- 
tually seen Byers but had seen the motion picture of him once, received the in- 
formation and decided to drive to Asbury Park to look for him. On the way, 
35 miles from Asbury Park, he passed a man who had a definitely familiar ap- 
pearance; he turned around, looked again, and arrested John A. Byers. The 
only thing that he had to rely upon was his memory of the sound motion picture 
he had seen of Byers. He certainly did not expect to find him 35 miles from As- 
bury Park, but the impression left by the motion picture was so complete that, 
from a passing automobile, he made an identification and picked up the man he 
was looking for. 

(Following the paper, sound motion pictures of male and female actor-criminals 
going through the required technic were shown.) 



Feb., 1937] 



NEW MOTION PICTURE APPARATUS 



217 



DISCUSSION 

MR. RICHARDSON: I wonder whether, when the criminal realizes what is going 
on, he will act very naturally. 

MR. FRANK: I think so. There is some feeling on the part of law enforce- 
ment officers that so many persons want to get into the movies that they will 
commit crimes just to have their pictures taken. Furthermore, law enforcement 
officers tell me that they have pretty good control of the criminals, and that 
taking the pictures at different times, as suggested, will doubtless overcome the 
difficulty. 

MR. ELLIS: Is monitoring necessary? If so, I suppose a second person would 
be required for handling the equipment. 

MR. FRANK: No; we do not require monitoring. 

MR. PRESGRAVE: Why is a double system necessary? 

MR. FRANK: The only reason why the double system has been used is that 
we did not have a single system available employing a ground-noise reduction 
system. Because of the high quality of speech required, that is the only system 
that has been used. 



COMMITTEES 

of the 
SOCIETY OF MOTION PICTURE ENGINEERS 

(Correct to January 20, 1936; additional appointments or changes may be 
made at any time during the year as necessity or expediency may require) 



L. W. DAVEE 
A. S. DICKINSON 



ADMISSIONS 

T. E. SHEA, Chairman 
M. W. PALMER 
H. RUBIN 



H. GRIFFIN 
D. E. HYNDMAN 



A. N. GOLDSMITH 
A. C. HARDY 



BOARD OF EDITORS 

J. I. CRABTREE, Chairman 

L. A. JONES 
E. W. KELLOGG 



H. G. KNOX 
T. E. SHEA 



W. H. CARSON 
O. O. CECCARINI 



COLOR 

J. A. BALL, Chairman 
C. H. DUNNING 
R. M. EVANS 

A. M. GUNDELFINGER 



H. W. MOYSB 
A. WARMISHAM 



H. GRIFFIN 



H. BUSCH 

A. S. DICKINSON 

G. C. EDWARDS 



CONVENTION 
W. C. KUNZMANN, Chairman 

J. H. KURLANDER 
EXCHANGE PRACTICE 

T. FAULKNER, Chairman 
A. HIATT 
J. S. MACLEOD 



M. W. PALMER 



N. F. OAKLEY 
H. RUBIN 
J. H. SPRAY 



J. E. ABBOTT 
T. ARMAT 



A. N. GOLDSMITH 
A. C. HARDY 



HISTORICAL 

E. THEISEN, Chairman 
G. A. CHAMBERS 
W. CLARK 

HONORARY MEMBERSHIP 

J. G. FRAYNE, Chairman 



G. E. MATTHEWS 
T. RAMSAYE 



H. G. TASKER 
W. E. THEISEN 



218 



COMMITTEES OF THE SOCIETY 



219 



E. HUSB 

K. F. MORGAN 



JOURNAL AWARD 

A. C. HARDY, Chairman 



G. F. RACKETT 
E. A. WILLIFORD 



LABORATORY PRACTICE 

D. E. HYNDMAN, Chairman 

J. CRABTREE M. S. LESHING H. W. MOYSE 

R. M. EVANS C. L. LOOTENS J. M. NICKOLAUS 

E. HUSE R. F. MITCHELL W. A. SCHMIDT 

T. M. INGMAN J. H. SPRAY 



MEMBERSHIP AND SUBSCRIPTION 

E. R. GEIB. Chairman 



Atlanta 

C. D. PORTER 

Boston 

T. C. BARROWS 
J. R. CAMERON 
J. S. CIFRE 

Camden & Philadelphia 
H. BLUMBERG 
J. FRANK, JR. 

Chicago 
B. W. DEPUE 
J. H. GOLDBERG 
S. A. LUKES 
R. F. MITCHELL 



Cleveland 
R. E. FARNHAM 
J. T. FLANNAGAN 
V. A. WELMAN 

Hollywood 
J. O. AALBERG 
L. E. CLARK 
G. H. GIBSON 
C. W. HANDLEY 

E. HUSE 

F. E. JAMES 

G. A. MITCHELL 
P. MOLE 

K. F. MORGAN 
G. F. RACKETT 

Minneapolis 
C. L. GREENE 



New York 

G. C. EDWARDS 

J. J. FINN 

G. P. FOUTE 

H. GRIFFIN 

W. W. HENNESSEY 

R. C. HOLSLAG 

M. D. O'BRIEN 

F. H. RICHARDSON 

H. B. SANTEE 

T. E. SHEA 

J. L. SPENCB 

J. H. SPRAY 

Rochester 

E. K. CARVER 

Washington 
N. GLASSER 

F. J. STORTY 



Australia 
H. C. PARISH 

Austria 

P. R. VON SCHROTT 

China 

R. E. O'BOLGER 



Canada 

F. C. BADGLEY 

C. A. DENTELBECK 

G. E. PATTON 



England 

W. F. GARLING 

R. G. LlNDERMAN 
D. MCMASTER 

R. TERRANEAU 
S. S. A. WATKINS 



220 



COMMITTEES OF THE SOCIETY 



[J. S. M. P. E. 



France 
L. J. DIDIEE 
L. G. EGROT 
F. H. HOTCHKISS 
J. MARETTE 

Germany 

W. F. BIELICKE 

K. NORDEN 

Hawaii 

L. LACHAPELLE 



O. B. DEPUE 



D. P. BEAN 
F. E. CARLSON 
W. B. COOK 
H. A. DBVRY 



India 

H. S. MEHTA 
M. L. MISTRY 
M. B. PATEL 

Japan 

T. NAGASE 

Y. OSAWA 

New Zealand 
C. BANKS 

MUSEUM 

(Western) 

E. THEISEN, Chairman 



NON-THEATRICAL EQUIPMENT 
R. F. MITCHELL, Chairman 
E. C. FRITTS 
H. GRIFFIN 
R. C. HOLSLAG 



Russia 

A. F. CHORINE 

E. G. JACHONTOW 

Travelling 

E. AUGER 

K. BRENKERT 

W. C. KUNZMANN 

D. McRAB 

O. F. NEU 

H. H. STRONG 



A. REEVES 



J. H. KURLANDER 

E. Ross 
A. SHAPIRO 
A. F. VICTOR 



L. N. BUSCH 
A. A. COOK 
L. J. J. DIDIEE 
M. E. GILLETTE 



PAPERS 
G. E. MATTHEWS, Chairman 

E. W. KELLOGG T. E. SHEA 

R. F. MITCHELL P. R. VON SCHROTT 

W. A. MUELLER H. C. SILENT 



H. B. SANTEE 



I. D. WRATTEN 



J. I. CRABTREB 
A. S. DICKINSON 
R. EVANS 



PRESERVATION OF FILM 
J. G. BRADLEY, Chairman 
M. E. GILLETTE 
C. L. GREGORY 



T. RAMSAYE 
V. B. SEASE 
W. A. SCHMIDT 



L. N. BUSCH 
G. A. CHAMBERS 
A. A. COOK 



PROGRESS 

J. G. FRAYNE, Chairman 
R. M. CORBIN 
R. E. FARNHAM 

A. GUNDELFINGER 

G. E. MATTHEWS 



H. MEYER 
V. E. MILLER 

G. WORRALL 



M. C. BATSEL 
J. I. CRABTREE 



PROGRESS AWARD 

A. N. GOLDSMITH, Chairman 



C. DREHER 
J. G. FRAYNE 



Feb., 1937] 



COMMITTEES OF THE SOCIETY 



221 



J. O. BAKBR 
T. C. BARROWS 

F. E. CAHILL 
J. R. CAMERON 

G. C. EDWARDS 
J. K. ELDERKIN 



PROJECTION PRACTICE 

H. RUBIN, Chairman 
J. J. FINN 

B. R. GEIB 

A. N. GOLDSMITH 
H. GRIFFIN 
J. J. HOPKINS 

C. F. HORSTMAN 
P. A. McGuiRE 



R. MlEHLING 

E. R. MORIN 

M. D. O'BRIEN 

F. H. RICHARDSON 
J. S. WARD 

V. WBLMAN 



PROJECTION SCREEN BRIGHTNESS 



A. A. COOK 
A. C. DOWNES 
D. E. HYNDMAN 



C. TUTTLE, Chairman 
W. F. LITTLB 
O. E. MILLER 
G. F. RACKETT 
H. RUBIN 



B. SCHLANGER 

A. T. WILLIAMS 
S. K. WOLF 



J. R. CAMERON 
J. J. FINN 



PUBLICITY 

W. WHITMORE, Chairman 
G. E. MATTHEWS 



P. A. McGuiRB 
F. H. RICHARDSON 



M. C. BATSEL 
L. E. CLARK 
F. J. GRIGNON 



SOUND 

P. H. EVANS, Chairman 
K. F. MORGAN 
O. SANDVIK 
E. I. SPONABLE 



R. O. STROCK 
H. G. TASKER 
S. K. WOLF 



P. ARNOLD 
F. C. BADGLEY 
M. C. BATSEL 
L. N. BUSCH 
W. H. CARSON 
A. CHORINE 
A. COTTET 
L. DE FEO 
A. C. DOWNES 
J. A. DUBRAY 



STANDARDS 

E. K. CARVER, Chairman 
P. H. EVANS 
R. E. FARNHAM 
C. L. FARRAND 
G. FRIEDL, JR. 
H. GRIFFIN 
R. C. HUBBARD 
E. HUSE 
C. L. LOOTENS 
W. A. MACNAIR 
K. F. MORGAN 
T. NAGASE 



N. F. OAKLEY 
G. F. RACKETT 
W. B. RAYTON 
C. N. REIFSTECK 
H. RUBIN 

0. SANDVIK 
H. B. SANTEE 
J. L. SPENCE 
A. G. WISE 

1. D. WRATTEN 



W. C. KUNZMANN 

J. H. KURLANDER 



STUDIO LIGHTING 

R. E. FARNHAM, Chairman 
V. E. MILLER 
G. F. RACKBTT 



E. C. RICHARDSON 

F. WALLER 



222 COMMITTEES OF THE SOCIETY 

SECTIONS OF THE SOCIETY 
(Atlantic Coast) 

G. FRIEDL, JR., Chairman 

L. W. DAVEE, Past-Chairman M. C. BATSEL, Manager 

D. E. HYNDMAN, Sec.-Treas. H. GRIFFIN, Manager 

(Mid-West) 

C. H. STONE, Chairman 

R. F. MITCHELL, Past-Chairman O. B. DEPUE, Manager 

S. A. LUKES, Sec.-Treas. B. E. STECHBART, Manager 

(Pacific Coast) 
K. F. MORGAN, Chairman 

H. G. TASKER, Past-Chairman J. O. AALAERG, Manager 

G. A. CHAMBERS, Sec.-Treas. H. W. MOYSE, Manager 



SPRING, 1937, CONVENTION 
SOCIETY OF MOTION PICTURE ENGINEERS 

HOLLYWOOD-ROOSEVELT HOTEL 
HOLLYWOOD, CALIF. 
MAY 24th-28th, INCL. 

Officers and Committees in Charge 

W. C. KUNZMANN, Convention Vice-President 

J. I. CRABTREE, Editorial Vice-President 

H. G. TASKER, Past-President 

G. F. RACKBTT, Executive Vice-President 

K. F. MORGAN, Chairman, Pacific Coast Section 

G. E. MATTHEWS, Chairman, Papers Committee 

Local Arrangements and Reception Committee 

P. MOLE, Chairman 

E. HUSE K. F. MORGAN H. G. TASKER 

J. O. AALBERG C. W. HANDLEY G. F. RACKETT 

C. N. BATSEL R. H. McCuLLOUGH H. W. MOYSE 

G. A. CHAMBERS G. S. MITCHELL W. J. QUINLAN 

Information and Registration 

W. C. KUNZMANN, Chairman 
E. R. GEIB S. HARRIS C. W. HANDLEY 

Ladies' Reception Committee 

MRS. K. F. MORGAN and MRS. P. MOLE, Hostesses 

assisted by 

MRS. F. C. COAXES MRS. E. HUSE MRS. E. C. RICHARDSON 

MRS. C. W. HANDLEY MRS. W. J. QUINLAN MRS. H. G. TASKER 
MRS. G. F. RACKETT 

Banquet 

E. HUSE, Chairman 

W. C. KUNZMANN K. F. MORGAN G. F. RACKETT 

P. MOLE W. J. QUINLAN H. G. TASKER 



224 SPRING CONVENTION [J. S. M. P. E. 

Projection Committee 

H. GRIFFIN, Chairman 

J. O. AALBERG J. G. FRAYNE C. W. HANDLEY 

L. E. CLARK G. M. GROSJEAN R. H. McCuLLOUGH 

Officers and Members of Los Angeles Local No. 150, I.A.T.S.E. 

Hotel Accommodations Committee 

G. F. RACKETT, Chairman 
E. HUSE K. F. MORGAN 

W. C. KUNZMANN H. G. TASKER 

H. C. SILENT 

Transportation Committee 

C. W. HANDLEY, Chairman 
G. A. CHAMBERS S. HARRIS 

H. GRIFFIN F. E. JAMES 

Publicity 

W. WHITMORE, Chairman 
J. J. FINN S. HARRIS 

W. GREENE G. E. MATTHEWS 

W. A. MUELLER 

Membership 

E. R. GEIB, Chairman 
G. A. CHAMBERS W. GREENE S. HARRIS 

Headquarters 

Headquarters of the Convention will be the Hollywood-Roosevelt Hotel, where 
excellent accommodations are assured. A reception suite will be provided for the 
Ladies' Committee, and an excellent program of entertainment will be arranged 
for the ladies who attend the Convention. 

Special hotel rates, guaranteed to SMPE delegates, European plan, will be as 
follows: 

One person, room and bath $3 . 50 

Two persons, double bed and bath 5.00 
Two persons, twin beds and bath 6 . 00 
Parlor suite and bath, 1 person 9 . 00 

Parlor suite and bath, 2 persons 12 . 00 

Room reservation cards will be mailed to the membership of the Society in the 
near future, and everyone who plans to attend the Convention should return his 
card to the Hotel promptly in order to be assured of satisfactory accommodations. 
Special garage rates will be provided for SMPE delegates who motor to the 
Convention. 



Feb., 1937] SPRING CONVENTION 225 

Railroad Fares 

The dates of the Convention have been chosen in order that delegates may avail 
themselves of the summer tourists' rates, which go into effect May 15th. The 
following table lists the railroad fares and Pullman charges: 

Railroad 

Fare Pullman 

City (round trip) (one way) 

Washington $ 120 . 75 $20 . 50 

Chicago 86.00 15.75 

Boston 132.80 22.25 

Detroit 98.30 18.00 

New York 126.90 21.75 

Rochester 112.50 19.25 

Cleveland 101.35 18.00 

Philadelphia 122 .85 21 . 00 

Pittsburgh 107.10 18.75 

The railroad fares given above are for round trips, forty-five day limits. Ar- 
rangements may be made with the railroads to take different routes going and 
coming, if so desired, but once the choice is made it must be adhered to, as changes 
in the itinerary may be effected only with considerable difficulty and formality. 
Delegates should consult their local passenger agents as to schedules, rates, and 
stop-over privileges. 

New streamlined trains will be operating from Chicago to Los Angeles and San 
Francisco, making the trip to Los Angeles in 39 hours. Special fares are levied on 
these trains. 

Technical Sessions 

An attractive program of technical papers and presentations is being arranged 
by the Papers Committee. Several sessions will be held in the evening, to permit 
those to attend who would be otherwise engaged in the daytime. All sessions will 
be held at the Hotel. 

Semi-Annual Banquet 

The Semi- Annual Banquet of the Society will be held at the Hotel on Wednes- 
day, May 26th. Addresses will be delivered by prominent members of the indus- 
try, followed by dancing and entertainment. Tables reserved for 8, 10, or 12 per- 
sons; tickets obtainable at the registration desk. 

Inspection Tours and Diversions 

Arrangements are under way to visit one or more of the prominent Hollywood 
studios, and passes will be available to registered members to several Hollywood 
motion picture theaters. Arrangements may be made for golfing and for special 
trips to points of interest in and about Hollywood. 



226 SPRING CONVENTION [j. s. M. p. E. 

Ladies' Program 

An especially attractive program for the ladies attending the Convention is be- 
ing arranged by Mrs. K. F. Morgan and Mrs. P. Mole, hostesses, and their Ladies' 
Committee. A suite will be provided in the Hotel, where the ladies will register 
and meet for the various events upon their program. Further details will be pub- 
lished in a succeeding issue of the JOURNAL. 

Points of Interest 

En route: Boulder Dam, Las Vegas, Nevada; and the various National Parks. 
Hollywood and vicinity: Beautiful Catalina Island; Zeiss Planetarium; Mt. 
Wilson Observatory; Lookout Point, on Lookout Mountain; Huntington Li- 
brary and Art Gallery (by appointment only) ; Palm Springs, Calif. ; Beaches at 
Ocean Park and Venice, Calif.; famous old Spanish missions; Los Angeles Mu- 
seum (housing the SMPE motion picture exhibit); Mexican village and street, 
Los Angeles. 

TENTATIVE PROGRAM 
Monday, May 24th 
10:00 a.m. Blossom Room 
Registration 
Society Business 
Committee Reports 
Technical Papers Program 
12:30 p.m. Florentine Room 

Informal Get-Together Luncheon for members, their families, and 
guests. Brief addresses by several prominent members of the 
industry. 
2:00 p.m. Blossom Room 

Technical Papers Program. 
8:00 p.m. (To be announced later.) 

Tuesday, May 25th 

10:00 a.m. Blossom Room 

Technical Papers Program 
2:00 p.m. (To be announced later.) 
8:00 p.m. Blossom Room 

Technical Papers Program 

Wednesday, May 26th 

10:00 a.m. Blossom Room 

Technical Papers Program 
2:00 p.m. (To be announced later.) 
8:00 p.m. Blossom Room 

Semi- Annual Banquet and Dance of the SMPE; Addresses by 
eminent members of the industry; dancing and entertainment. 



Feb., 1937] SPRING CONVENTION 227 

Thursday, May 27th 

10:00 a.m. Open morning 
2:00 p.m. Blossom Room 

Technical Papers Program 
8:00 p.m. Blossom Room 

Technical Papers Program 

Friday, May 28th 

10:00 a.m. Blossom Room 

Technical Papers Program 
2:00 p.m. Blossom Room 

Technical Papers Program 

Open Forum 

Adjournment of the Convention 

NOTE: All technical sessions will be held in the Blossom 
Room of the Hollywood-Roosevelt Hotel. There will be no 
public exhibit of apparatus anywhere in the Hotel; although 
members registered in the Hotel will, of course, be privileged 
to display any equipment they wish in their own rooms. 



SOCIETY ANNOUNCEMENTS 
BOARD OF GOVERNORS 



Further plans for the approaching Hollywood Convention in May were for- 
mulated by the Board of Governors at a meeting held in the Hotel Pennsylvania, 
New York, N. Y., on January 7th. In addition, the budget for 1937 and various 
other administrative matters were completed. The Financial Vice-President re- 
ported favorable performance during the past year, and that the membership 
of the Society had attained a new all-time high, the total paid membership at 
the end of the year numbering 1230. 

The new West-Coast Office of the Society, located in Suite 226 of the Equitable 
Building, Hollywood, has begun to function, under the charge of Mr. Walter 
Greene, and a European Advisory Committee has been established under the 
Chairmanship of Mr. J. Van Breukelen, of Eindhoven, Holland. The other 
members of the European Committee are: F. H. Hotchkiss, of Paris; H. Warncke, 
of Berlin; and I. D. Wratten, of London. 

As a result of the recent election, the Officers and Managers of the West-Coast 
Section for the year 1937 are as follows: 



Chairman: K. F. Morgan 
Past-Chairman: G. F. Rackett 
Sec.-Treas.: G. A. Chambers 
Governors: J. O. Aalberg 
H. W. Moyse 



In the recent elections for officers of the national Society for 1937, Mr. H. G. 
Tasker, who automatically retained membership upon the Board of Governors 
in the capacity of Past-President, was elected Executive Vice-President of the 
national Society. In view of the fact that he would be holding two offices con- 
currently, Mr. Tasker submitted to the Board his resignation as Executive Vice- 
President. By unanimous vote Mr. G. F. Rackett was elected by the Board to 
serve as Executive Vice-President for 1937. 

The complete list of Officers and Members of the Board of Governors of the 
Society will be found upon the reverse of the contents page of this and sub- 
sequent issues of the JOURNAL, and the personnel of the Boards of Managers 
of the Local Sections following the list of technical committees, also in each 
issue. 

228 



SOCIETY ANNOUNCEMENTS 



229 



ADMISSIONS COMMITTEE 

At a recent meeting of the Admissions Committee, at the General Office of the 
Society, the following applicants for membership were admitted to the Associate 
grade: 

ABBOTT, L. R. HILTON, R. G. 

2323 N. Kostner Ave. 36-08 146th St. 

Chicago, 111. Flushing, N. Y. 



ANDERSON, J. C. 
6114 Stanton Ave. 
Pittsburgh, Pa. 

BANG, P. 

Kodak Aktieselskab 
Vodroffsvej 26 
Copenhagen V, Denmark 

BINDER, O. C. 

Universal Pictures Corp. 
1250 Sixth Ave. 
New York, N. Y. 

BLACK, J. G. 
3333 N. Marshfield Ave. 
Chicago, 111. 

BRIGHTON, A. F. 
913 E. Kilbourn Ave. 
Milwaukee, Wis. 

CHALLENNER, A. 
3139 N. W. 14th St. 
Oklahoma City, Okla. 

CLARK, L. M. 

187 Westminster Road 
Rochester, N. Y. 

DAY, G. 

701 Franklin Ave. 
Brooklyn, N. Y. 

DEWEY, D. H. 
300 Capitol Theater Bldg. 
Des Moines, Iowa 

FROSCH, M. E. 
56 Glenwood Ave. 
Minneapolis, Minn. 

GALLAGHER, L. W. 
7602 E. Lake Terrace 
Chicago, 111. 



KEEP, N. 
599 Eighth St. 

San Francisco, Calif. 

MARTIN, P., JR. 

U. S. Soldiers Home 
Washington, D. C. 

PALMER, W. A. 
369 Churchill Ave. 
Palo Alto, Calif . 

SCHUYLER, J. B. 

3925 N. Downer Ave. 
Milwaukee, Wis. 

SENNETT, P. T. 
Tiffin Scenic Studios 
Tiffin, Ohio 

SHARP, T. C. 
4379 Camellia 

North Hollywood, Calif. 

SHOLKIN, A. N. 

RCA Communications, Inc. 
66 Broad St. 
New York, N. Y. 

SKIMIN, G. J. 
542 Neff Road 

Grosse Pointe, Mich. 

STIMSON, S. 
2236 82d St. 

Brooklyn, N. Y. 

WETTER, R. 
40 Piedmont St. 
Boston, Mass. 

WILLEY, L. E. 

717 N. Ramage St. 

West Hollywood, Calif. 



230 SOCIETY ANNOUNCEMENTS [J. S. M. P. E. 

In addition, the following applicants have been admitted by vote of the Board 
of Governors to the Fellow and Active grades: 
ARNOLD, E. L. (M) HARTMANN, G. (M) 

Eastman Kodak Co. 7130 La Presa Drive 

Kodak Park B/26 Hollywood, Calif. 

Rochester, N. Y. 

BUB, G. L. (M) OSWALD, C. L. (M) 

Building 3 433 W. 21st St. 

Second & Arsenal Sts. New York, N. Y. 
St. Louis, Mo. 

GOLDEN, N. D. (M) SHIMEK, J. A. (M) 

Bureau of Foreign and American Film Corp. 

Domestic Commerce 6227 Broadway 

U. S. Government Chicago, 111. 
Washington, D. C. 

GREEN, W. E. (F) TRIVELLI, A. P. H. (M) 

National Theater Supply Co. Eastman Kodak Co. 

92 Gold St. Research Laboratories 

New York, N. Y. Rochester, N. Y. 

INTERNATIONAL EXHIBITION OF APPLIED AND SCIENTIFIC 
PHOTOGRAPHY 

ROCHESTER, MARCH, 1937 

An International Exhibition of Applied and Scientific Photography will be held 
at Rochester in March, 1937, under the sponsorship of the Rochester Scientific 
and Technical Section of the Photographic Society of America. The objective of 
the exhibition will be to show examples of the application of photography to the 
various branches of science and technology. 

The following sections have been organized : 

(1) Color photography: (a) processes in detail, (b) transparencies, (c) prints. 

(2) Astronomy and metrology. 

(3) Aerial photography. 

(4) Photomicrography: (a) metallography, (b) other subjects. 

(5) Medical photography: (a) prints, (b) radiographs, (c) motion pictures. 

(6) X-Ray in industry. 

(7) Documentary photography: (a) small-film library work, (b) instrument 
reading, (c) miscellaneous. 

(8) High-speed photography. 

(9) Stereo-photography: (a) prints, (b) transparencies, (c) motion pictures. 

(10) Photography in physics and chemistry: (a) x-ray spectrography, (b) 
cosmic and other ray effects, (c) miscellaneous. 

(11) Photographic sensitivity: (a) photographic effects, (b) light-sensitive 
substances. 

(12) Natural history. 

(13) Miscellaneous. 



Feb., 1937] SOCIETY ANNOUNCEMENTS 231 

Photographs or apparatus showing the applications of photography to typical 
problems in any branch of science and technology will be welcomed. All corre- 
spondence in regard to the exhibition, or requests for entry blanks should be ad- 
dressed to the Secretary, C. B. Neblette, F.R.P.S., Department of Photographic 
Technology, Rochester Athenaeum and Mechanics Institute, Rochester, N. Y. 

COURSE IN PROJECTION PRACTICE 

Members of the IATSE from various locals of Florida attended a short course in 
projection practice and sound theory at the University of Florida School of 
Adult Education at Camp Roosevelt, the week of December 14th, with more 
than forty registered projectionists in attendance. 

The School of Adult Education, a new branch of the University of Florida, is 
conducting a broad program in training adults, utilizing the facilities of the base 
camp of the Atlantic-Gulf Ship Canal. Major B. C. Riley, dean of the General 
Extension Division, is in charge of the program. This short course for projec- 
tionists is the first of a series that the University School of Adult Education ex- 
pects to conduct this year. 

Prominent representatives from the leading motion picture sound and projec- 
tion apparatus manufacturers attended, and conducted classes in the proper 
operation and maintenance of their products. Equipment displayed and demon- 
strated was of the latest design and incorporated many features yet not on the 
market. The manufacturers spared no expense in cooperating with University 
officials in making the venture a success. 

It is believed that this is the first time that a study course of this type has been 
offered to IATSE projectionists. R. J. Gavin of local 511, Jacksonville, and 
George E. Raywood of local 316, Miami, assisted the University with arrange- 
ments. 

SOCIETY SUPPLIES 

The following are available from the General Office of the Society, at the prices 
noted. Orders should be accompanied by remittances. 

Aims and Accomplishments. An index of the Transactions from October, 
1916, to December, 1929, containing summaries of all articles, and author and 
classified indexes. One dollar each. 

Journal Index. An index of the JOURNAL from January, 1930, to December, 
1935, containing author and classified indexes. One dollar each. 

SMPE Standards. Reprints of SMPE Standards and Recommended Practice. 
Twenty-five cents each. 

Membership Certificates. Engrossed, for framing, containing member's name, 
grade of membership, and date of admission. One dollar each. 

Lapel Buttons. The insignia of the Society, gold filled, with safety screw back. 
One dollar each. 

Journal Binders. Black fabrikoid binders, lettered in gold, holding a year's 
issue of the JOURNAL. Two dollars each. Member's name and the volume 
number lettered in gold upon the backbone at an additional charge of fifty cents 
each. 

Test-Reels. See advertisement on following page. 



STANDARD S. M. P. E. 

VISUAL AND SOUND TEST REELS 

Prepared under the Supervision 

OF THE 
PROJECTION PRACTICE COMMITTEE 

OF THE 
SOCIETY OF MOTION PICTURE ENGINEERS 



Two reels, each approximately 500 feet long, of specially pre- 
pared film, designed to be used as a precision instrument in 
theaters, review rooms, exchanges, laboratories, and the like 
for testing the performance of projectors. The visual section 
includes special targets with the aid of which travel-ghost, 
lens aberration, definition, and film weave may be detected 
and corrected. The sound section includes recordings of 
various kinds of music and voice, in addition to constant- 
frequency, constant-amplitude recordings which may be used 
for testing the quality of reproduction, the frequency range 
of the reproducer, the presence of flutter and 60-cycle or 96- 
cycle modulation, and the adjustment of the sound-track. 
Reels sold complete only (no short sections). 

PRICE $37.50 FOR EACH SECTION, 
INCLUDING INSTRUCTIONS 

(Shipped to any point in the United States) 

Address the 

SOCIETY OF MOTION PICTURE ENGINEERS 

HOTEL PENNSYLVANIA 
NEW YORK, N. Y. 



I JOURNAL 

OF THE SOCIETY OF 

MOTION PICTURE ENGINEERS 



Volume XXVIU MARCH, 1937 Numbers 



CONTENTS 

Page 

The Influence of Sprocket Holes upon the Development of Adja- 
cent Sound-Track Areas. . J. G. FRAYNE AND V. PAGLIARULO 235 

Modern Theater Loud Speakers and Their Development 

C. FLANNAGAN, R. WOLF, AND W. C. JONES 246 

Research Council Specifications for a Standard Synchronizing 
System for Cameras 265 

Transmission of Sound and Vibration in Buildings 

E. MEYER 271 

The Use of Visual Equipment in Elementary and Secondary 
Schools C. M. KOON 284 

Organization and Work of the Film Library of the Museum of 
Modern Art J. E. ABBOTT 294 

A Neon-Tube Oscilloscope for the Projection Room 

F. H. RICHARDSON AND T. P. HOVER 304 

New Motion Picture Apparatus 

New Recording Equipment D. CANADY 309 

A New Reel-End Alarm. . . .D. CANADY AND V. A. WELMAN 314 
A Demonstration Triode for Visualizing Electronic Phenomena 

F. E. ELDREDGE AND H. F. DART 318 

Committees of the Society 323 

Spring, 1937, Convention: Hollywood, Calif., May 24th-28th, 
Inclusive 328 

Society Announcements 333 



JOURNAL 

OF THE SOCIETY OF 

MOTION PICTURE ENGINEERS 



SYLVAN HARRIS, EDITOR 

Board of Editors 
J. I. CRABTREE, Chairman 

A. N. GOLDSMITH L. A. JONES H. G. KNOX 

A. C. HARDY E. W. KELLOGG T. E. SHEA 



Subscription to non-members, $8.00 per annum ; to members, $5.00 per annum, 
included in their annual membership dues; single copies, $1.00. A discount 
on subscriptions or single copies of 15 per cent is allowed to accredited agencies. 
Order from the Society of Motion Picture Engineers, Inc., 20th and Northampton 
Sts., Easton, Pa., or Hotel Pennsylvania, New York, N. Y. 
Published monthly at Easton, Pa., by the Society of Motion Picture Engineers. 

Publication Office, 20th & Northampton Sts., Easton, Pa. 
General and Editorial Office, Hotel Pennsylvania, New York, N. Y. 

West-Coast Office, Suite 226, Equitable Bldg., Hollywood, Calif. 
Entered as second class matter January 15, 1930, at the Post Office at Easton, 
Pa., under the Act of March 3, 1879. Copyrighted, 1937, by the Society 
Motion Picture Engineers, Inc. 

Papers appearing in this Journal may be reprinted, abstracted, or abridged 
provided credit is given to the Journal of the Society of Motion Picture Engineers 
and to the author, or authors, of the papers in question. Exact reference as to 
the volume, number, and page of the Journal must be given. The Society is 
not responsible for statements made by authors. 



OFFICERS OF THE SOCIETY 

President: S. K. WOLF, 250 W. 57th St., New York, N. Y. 
Past-President: H. G. TASKER, Universal City, Calif. 

Executive Vice-P resident, G. F. RACKETT, 823 N. Seward St., Hollywood, Calif. 
Engineering Vice-President: L. A. JONES, Kodak Park, Rochester, N. Y. 
Editorial Vice-President: J. I. CRABTREE, Kodak Park, Rochester, N. Y. 
Financial Vice-President: O. M. GLUNT, 463 West St., New York, N. Y. 
Convention Vice-President: W. C. KUNZMANN, Box 6087, Cleveland, Ohio. 
Secretary: J. FRANK, JR., 90 Gold St., New York, N. Y. 
Treasurer: L. W. DAVEE, 250 W. 57th St., New York, N. Y. 

GOV NORS 

M. C. BATSEL, Front and Market Sts., Camden, N. J. 

A. S. DICKINSON, 28 W. 44th St., New York, N. Y. 

G. FRIEDL, JR., 250 W. 57th St., New York, N. Y. 

A. N. GOLDSMITH, 444 Madison Ave., New York, N. Y. 

H. GRIFFIN, 90 Gold St., New York, N. Y. 

A. C. HARDY, Massachusetts Institute of Technology, Cambridge, Mass. 

K. F. MORGAN, 7046 Hollywood Blvd., Los Angeles, Calif. 

C. H. STONE, 205 W. Wacker Drive, Chicago, 111. 






See p. 323 for Technical Committees 



THE INFLUENCE OF SPROCKET HOLES UPON THE 
DEVELOPMENT OF ADJACENT SOUND-TRACK AREAS* 

J. G. FRAYNE AND V. PAGLIARULO** 

Summary. An unmodulated sound-track shows 96-cycle modulation on develop- 
ment. The effect is a maximum at the edge of the sprocket holes and diminishes ex- 
ponentially for a distance of approximately 30 mils into the sound-track. A film 
modulated by a constant frequency shows 96-cycle amplitude and frequency modulation 
over the same area. Both effects are introduced principally during processing of the 
film. A film having no sprocket holes on the sound-track side is entirely free of these 
effects. The conclusion is that processing standards in many laboratories require 
improvement to eliminate distortions of this type. 

It has been generally understood for at least the past three years 
that uneven development of motion picture film occurs in the neigh- 
borhood of the sprocket holes, the effect being usually referred to as 
sprocket-hole modulation. 1 While the modulation has usually been 
considered as of the amplitude type, studies of flutter in film-pulling 
mechanisms has led to the belief that frequency modulation was 
also induced in the neighborhood of the sprocket holes. This was 
first noticed when the flutter or frequency modulation induced by 
sprocket teeth in a film recorder appeared to vary with the laboratory 
chosen to develop the negative film, and also varied from day to day 
in any particular laboratory. Also any change of location of the 
scanned image appeared to result in a change in the amount of flutter 
present in the record. 

In order to investigate the nature of these development irregulari- 
ties, a series of exposures, both unmodulated and modulated with a 
frequency of 3000 cps., were made on a single roll of variable-density 
sound-track negative film. The recording machine used for this 
work was specially selected because of its freedom from the 96-cycle 
modulation commonly called flutter. The recorded film was then 
broken down, and a strip containing both unmodulated and 3000- 

* Received October 14, 1936; presented at the Fall, 1936, Meeting at Roches- 
| ter, N. Y. 

** Electrical Research Products, Inc., Hollywood, Calif. 

235 



236 



J. G. FRAYNE AND V. PAGLIARULO [J. S. M. P. E. 



cycle exposures was sent to each of six different laboratories. The 
negatives were developed, and prints were made and developed in 
routine manner. 



ANALYSIS OF UNMODULATED TRACK 



The unmodulated strips, negatives, and prints were scanned on a 
recording microphotometer furnished through the courtesy of Mr. 
Douglas Shearer of the MGM Sound Department. The important 
parts of the apparatus are shown in Fig. 1, and consist of a sound 




(4) Reproducer head 
(J5) Tone wheel 
(C) Moving platform 
(D) P.E.C. amplifier 
(E) Potentiometer 
(F) Amplifier 


(G) 
tfO 
(/) 

(/) 

TO 

(L) 



FIG. 1. Schematic arrangement of microphotometer: 

Band-pass filter 

Rectifier 

Milliameter 

Shutter 

Recording machine 

Film 



reproducing head, the light-beam of which is interrupted by a tone 
wheel at the rate of 400 times per second; a moving platform upon 
which the film to be scanned is placed; a photoelectric cell and as- 
sociated amplifier, to receive the light impulses and change them 
into electrical impulses the intensities of which are controlled by a 
potentiometer; an amplifier; a band-pass filter; and a low-impedance 
full-wave rectifier. The output of the rectifier is impressed upon a 
d-c. voltmeter mechanism of very low natural period, which moves a 
shutter located in the optical path of a special film-recording machine. 



Mar., 1937] SPROCKET HOLES AND SOUND-TRACK AREAS 



237 



The movement of the shutter is proportional to the change of trans- 
mission of the sample of film being scanned. The speeds of the scan- 
ning platform and the film in the recording machine upon which 
the record is made may be changed to obtain any amplification ratio 
desired. In the present case a magnification of about 1 : 3 was used. 
It was found that sprocket-hole modulation was evident in every neg- 
ative and print, although the magnitude of the effect varied consider- 



AMPLITUDE 
MODULATION 



CENTER OF BEAM 
FROM EDGE OF TRACK 
MILS 



AMPLITUDE 
MODULATION 
Vo 




FIG. 2. (Left) Typical amplitude modulation induced by sprocket holes in 

sound negative. 

FIG. 3. (Right) Typical amplitude modulation induced by sprocket holes in 

sound print. 

ably with the laboratory. The results shown in Figs. 2 and 3 are 
typical of sprocket-hole amplitude modulation found in all six 
laboratories, the former showing the effect upon a negative and the 
latter upon a print made from the negative, processed at the same 
laboratory. Reference to these sketches will show that the sound- 
track area was scanned in 5-mil widths beginning at the edge of the 
sound-track, the height of the scanning beam being adjusted to 
5 mils so as to provide sufficient output for satisfactory operation of 
the microphotometer. In every case the height of the dark portion 
of the graph is proportional to the transmission of the sample. Fig. 
2 indicates that the modulation induced by the presence of the 



238 



J. G. FRAYNE AND V. PAGLIARULO [J. S. M. P. E. 



sprocket holes diminished from 14.4 per cent next to the holes to 
0.4 per cent 27.5 mils from the holes. 

In order to determine the position of maximum development in 
the region of the sprocket holes, a fine black line was drawn upon 
the developed samples between the sprocket holes. Fig. 4 shows the 
result of scanning these samples. The graph for the negative shows 
that the opacity at this point lies near the trough of the wave cor- 
responding to the point of lowest transmission on the film. Con- 
versely, the crest of the wave on the record lies between the vertical 
lines forming the boundaries of the sprocket-hole area. This shows 



D D D D 








FIG. 4. Location of local development ac- 
tion around sprocket holes, for negative and 
print. 



that the density of the sound-track area was below average in the 
region opposite the sprocket hole, and above average in the region 
opposite the area between successive sprocket holes. Fig. 4 also shows 
that the effect is reversed in the print. In this case the maximum 
transmission occurred in the area opposite the region between sprocket 
holes. Further experiments indicated that sprocket-hole modulation 
in the print must be largely attributed to modulation in the negative 
from which the print is made, rather than to that inherent in the 
development of the print. This may be accounted for by the low 
gamma of approximately 0.35 of the negative development in a 
film stock having a gamma infinity of over 2.0. In such stock, local 



Mar., 1937] SPROCKET HOLES AND SOUND-TRACK AREAS 



239 



developing action, due to sprocket-hole turbulation in the solution, 
may produce a silver deposit corresponding to what might be obtained 
in a more normal development at a much higher gamma. On the 
print, the gamma of 2.10 used is so near the gamma infinity value 
that the effects of excess local development become minimized. 
The sprocket-hole modulation generally extended for 30 mils 






*6* T "t 






FIG. 5. Variation of per cent amplitude 
modulation with distance (in mils) from edge of 
sound-track near sprocket holes: () measured 
with microphotometer ; (x) measured with 
modified flutter set. 

into the sound area. The rate of decrease of the amplitude of the 
effect is shown for a laboratory, among those checked, in Fig. 5, 
the curves being of an exponential nature. The points indicated by 
circles were obtained from measurements made in a manner to be 
described later in this paper. The residual modulation indicated at a 
distance of 30 mils in from the sprocket holes might be attributed 
to the grain structure of the silver deposit, is generally higher in a 
print than in a negative, and is not typically 96 cycles. 



240 



J. G. FRAYNE AND V. PAGLIARULO [J. S. M. P. E. 

ANALYSIS OF A 3000-CYCLE RECORDING 



This analysis was made using a flutter-measuring equipment 2 
modified so that amplitude as well as frequency modulation could be 



CENTER OF BEAM FREQUENCY 

FROM EDGE OF TRACK MODULATION 



AMPLITUDE 
MODULATION 




AMPLIT. 
MOD.K> 



0.22 I.I 



0.16 1.0 



0. 1 1 0.0 



2.5 MILS 



17.5 MILS 



27.5 MILS 




0.41 



032 



0.27 



0.23 



FIG. 6. (Upper} Oscillograms of frequency and ampli- 
tude modulation from negative with sprocket holes near 
sound-track. Film modulated 3000 cps. 

FIG. 7. (Lower) Oscillograms of frequency and ampli- 
tude modulation from print with sprocket holes near sound- 
track. Film modulated 3000 cps. 

measured. In this analysis the procedure followed was that used in 
making ordinary flutter measurements, except that the scanning 
beam, normally 80 mils wide, was made adjustable. The output of 



Mar., 1937] SPROCKET HOLES AND SOUND-TRACK AREAS 



241 



the flutter equipment, after suitable amplification, was impressed 
upon a string of the rapid-record oscillograph, 8 and records made as 
shown in Figs. 6 and 7. The left-hand portion of these records in- 
dicates frequency modulation, the right-hand portion indicating 
amplitude modulation of the same piece of sound-track. When the 
per cent of amplitude modulation is plotted as a function of the dis- 
tance of the scanning beam from the sprocket-hole edge, the results 
shown by the dotted circles on the curves of Fig. 5 agree roughly 




20 30 40 50 60 70 

DISTANCE IN MILS FROM EDGE or SOUND-TRACK 

FIG. 8. Variation of per cent frequency flutter with distance (in 

mils) from edge of sound-track near sprocket holes. 
(A ) Measured and computed per cent frequency flutter for 80 mils 

scanning in print. 
() Measured per cent frequency flutter for 80 mils scanning in 

negative. 

(C) Computed per cent frequency flutter for 80 mils scanning in 

negative. 

with the curves obtained from the microphotometer measurements. 
The discrepancies may be partially attributed to the fact that 
in some cases the film zone that was scanned may have differed by 
a few mils in the two methods followed. 

Examination of the records in Figs. 6 and 7 shows that the fre- 
quency of both types of modulation is 96 cps., corresponding to the 
number of sprocket holes per second passing the scanning beam. 
It will be noted that as the scanning zone is moved away from the 
sprocket holes, the regular 96-cycle modulation is replaced by an 



242 



J. G. FRAYNE AND V. PAGLIARULO [J. S. M. P. E. 



CENTER OF BEAM 
FROM EDGE OT TRACK 



2.5 MILS 




FREQ. A, 'PLIT 

MOD. * M0i>.<*> 



009 I.I 



0.08 I.I 



7.5 MILS 




FREQ. AMPLIT. 

MOD. *> MOD. % 



FIG. 9. (Upper) (A) Microphotometer record of trans- 
mission in unmodulated sound negative without sprocket holes 
near sound-track. 

(B) Oscillograms of frequency and amplitude modulations 
from negative without sprocket holes near sound-track. Film 
modulated 3000 cps. 

FIG. 10. (Lower) (A) Microphotometer record of trans- 
mission in unmodulated sound print with sprocket holes near 
sound-track made from negative without sprocket holes. 

(B) Oscillograms of frequency and amplitude modulation 
from print with sprocket holes near sound-track made from 
negative without sprocket holes. Film modulated 3000 cps. 



Mar., 1937] SPROCKET HOLES AND SOUND-TRACK AREAS 243 

irregular one attributable without doubt to the random distribution 
of the silver grains in the film. 

The curves of Fig. 8 show the relation between frequency modula- 
tion and the distance from the edge of sprocket holes of the center of 
the 5-mil scanning zone, for both the negative and the print made 
from it at another laboratory. Taking the negative first, it will 
be noted that the flutter varies from 0.54 per cent for the 5-mil zone 

CENTER OF BEAM 
fROM EDGE Of TRACK A 



2. 5 MILS 



7.5 MILS 




FREQ AMPLIT 
MOO.-MJ MOD.* 



2. 5 MILS 0.23 



7. 5 MILS 



FIG. 11. (-4) Microphotometer record of transmis- 
sion in unmodulated sound print without sprocket holes 
near sound-track made from negative without sprocket 
holes. 

(B) Oscillograms of frequency and amplitude modu- 
lation from print without sprocket holes near sound- 
track made from negative without sprocket holes. 
Film modulated 3000 cps. 

adjacent to the sprocket holes, to 0.05 per cent when the zone is 
moved in 100 mils. Similarly the values for the print range from 
0.9 to 0.1 per cent. The usual width of sound-track scanned in 
reproducing is approximately 80 mils, and the center of this zone is 
54 mils removed from the sprocket-hole edge. Accordingly, vertical 
lines were drawn upon the graph marking the boundaries of the usual 
scanning area, and the intercepted areas lying beneath the negative 
and the positive curves were measured with a planimeter. The 
areas were then divided by the 80-mil base line, and the resulting 



244 J. G. FRAYNE AND V. PAGLIARULO [j. S. M. p. E. 

ordinates gave the average frequency modulation to be expected from 
these films. These values correspond very closely to the measured 
values obtained, respectively, for the negative and positive, and are 
shown upon the graph. 

This illustrates very clearly that when frequency modulation or 
flutter is measured in the standard manner, the magnitude of the 
quantity measured may be attributed in part to the development 
and in part to the recording or printing mechanism, as well as in part 
to random modulation. Thus, in the case under study, it is found 
that in the negative track at a distance of 100 mils from the sound- 
track edge, the flutter is reduced to a value of 0.05 per cent. Hence 
the recorder used in this test might be said to be essentially free of 
96-cycle flutter, because if flutter had been introduced by the machine 
it would have extended the entire width of the track. In the case of 
the print, flutter is higher throughout the region studied, reaching an 
apparent residual value of 0.10 or 0.05 per cent higher than the nega- 
tive. This indicates that but little 96-cycle flutter is introduced in 
printing, and the rise in residual flutter may be attributed to other 
causes. 

35-MM. FILM WITH ONE SET OF PERFORATIONS 

In order to determine more definitely the part played by the 
sprocket holes in producing the results analyzed above, the pos- 
sibility was suggested of obtaining 35-mm. film having perforations 
along one side only. Accordingly the Eastman Kodak Company very 
generously supplied 1000 feet of both sound negative and positive 
emulsions so perforated. In order to make the tests the teeth on the 
sound-track side were removed from the sprockets of the recorder 
to accommodate the negative stock during recording; and in order to 
make a print from the negative, the teeth were ground off one side 
of the sprockets of a standard printer and prints were made on film 
having (a) one set of perforations, (b) having two sets of perforations. 

The results obtained from the negative are shown in Fig. 9, both 
microphotometer and flutter set results being given. It will be noted 
that no 96-cycle frequency or amplitude modulation is in evidence. 
The results obtained from the prints are shown in Figs. 10 and 11, 
the top curves in Fig. 10 being microphotographs from the print 
having sprocket holes. These show some evidence of 96-cycle ampli- 
tude modulation. Incidentally, the same curves show how relatively 
unimportant this effect is in a print made from a negative from which 



Mar., 1937] SPROCKET HOLES AND SOUND-TRACK AREAS 245 

the effect was absent. In the lower part of Fig. 10 are the oscillograms 
of flutter measurements made from the 3000-cyde modulated print. 
Here frequency modulation is very pronounced. In Fig. 1 1 are given 
the graphs of the print that had no sprocket holes near the sound- 
track, showing the total absence of both amplitude and frequency 
modulation. The fact that the print having sprocket holes, but made 
from a negative without sprocket holes adjacent to the sound-track 
area, shows decided evidence of frequency modulation but little 
of amplitude modulation, while the other print having no sprocket 
holes made from the same negatives shows the absence of both modu- 
lations, points to the possibility that frequency modulation induced 
by the development process may be attributed to differential shrink- 
age of the film in the areas between and opposite sprocket holes. We 
have for some time past had experimental evidence, in another 
connection, to indicate that frequency modulation, to the extent of 
0.25 to 0.5 per cent, may be introduced by improperly drying nega- 
tives, and that properly hardening the film before drying very greatly 
reduces the hazard of distortion from this cause. 

The cause of amplitude modulation may be directly attributed 
to the passage of the sprocket holes through the developing solution. 
This is verified by the fact that when film without sprocket holes is 
used the phenomenon disappears. 

REFERENCES 

1 CRABTREE, J., AND WADDELL, J. H.: "Directional Effects in Sound-Filra 
Processing II," J. Soc. Mot. Pict. Eng., XXI (Nov., 1933), No. 5, p. 351. 

J SCOVILLE, R. R.: "A Portable Flutter-Measuring Instrument," /. Soc. 
Mot. Pict. Eng., XXV (Nov., 1935), No. 5, p. 416. 

SHEA, T. E., CURTIS. A. M., AND RUMPEL, C. H.: "The Rapid-Record Oscillo- 
graph in Sound Picture Studies," /. Soc. Mot. Pict. Eng. XVIH (Jan., 1932), 
No. 1. p. 39. 



MODERN THEATER LOUD SPEAKERS AND THEIR 
DEVELOPMENT* 



C. FLANNAGAN,** R. WOLF,** AND W. C. JONESf 



Summary. Although many of the basic ideas involved in the operation of present- 
day loud speakers were conceived during the early stages of the development of the 
telephone, it was not until the advent of the vacuum tube amplifier that these principles 
were applied to the design of structures capable of delivering sufficient acoustical 
power to be audible throughout a room or auditorium. Having reached this stage, 
however, the developments that culminated in the sound reproducing systems em- 
ployed with present-day sound pictures came in rapid succession. These develop- 
ments have embraced all phases of loud speaker design, with the result that systems 
are now available that convert from 25 to 50 per cent of the electrical input into acous- 
tical output, and maintain conversion efficiencies of this order of magnitude over a 
frequency range of 50 to 10,000 cps. These systems are so designed as to be capable 
of reproducing the recorded sound at intensities that not only greatly enhance the 
dramatic effect of the presentation in the theater, but also open entirely new fields 
in recording. All these improvements have been attained with a reduction in dis- 
tortion and improved fidelity of the reproduced sound. The directional properties 
of the loud speakers also have been markedly improved, with the result that the better 
quality of reproduction achieved is available throughout the entire seating area and 
the undesirable beam effects previously experienced have been eliminated. 

Almost without exception the basic ideas underlying the operation 
of the loud speaker were conceived during the early stages of develop- 
ment of electrical transmission of speech. For example, the magnetic 
circuit used in the motor elements of many of the first commercial 
loud speakers was employed also by Bell in his demonstration models. 
In this circuit the polarizing field in the air-gaps between fixed pole- 
pieces and a movable magnetic diaphragm is modulated by the field 
set up by the voice currents. Although it possesses such desirable 
characteristics as simplicity and cheapness, this magnetic circuit 
is poorly adapted to loud speakers designed to deliver large amounts 
of acoustical power, owing to the fact that the diaphragm represents 

* Received January 18, 1937; presented at the Fall, 1936, Meeting at Roches- 
ter, N. Y. 

** Electrical Research Products, Inc., New York, N. Y. 

t Bell Telephone Laboratories, Inc., New York, N. Y. 
246 



MODERN THEATER LOUD SPEAKERS 



247 



a compromise between the magnetic characteristics required at high 
efficiencies and the mechanical properties necessary for good response. 
Furthermore, a certain amount of distortion in the form of a double- 
frequency component is always present in the output. Although this 
is a relatively unimportant factor at low output intensities, it rapidly 
assumes a position of importance as the output level is increased. 

Means for correcting these limitations were proposed at an early 
date. As early as 1880 the suggestion was made that the performance 
of the telephone receiver could be improved by the use of the balanced- 
armature type of magnetic circuit. l Other disclosures along this line 
were made in rapid succession, with 
the result that by 1890 structures such 
as that shown in Fig. 1, were proposed 
embodying practically all the basic 
features of the balanced-armature 
type of magnetic circuit which has 
found such extensive use in loud 
speakers. 2 In addition to being in- 
herently more efficient than the simple 
form of magnetic circuit previously 
discussed, this circuit possesses also the 
desirable characteristics of materially 
reducing the double-frequency com- 
ponent in the output and enabling 
the designer to achieve high magnetic 
efficiency without adding unduly to the 
effective mass of the moving element. 
The output at low frequencies, how- 
ever, is limited by the non-linear distortion that occurs with large 
displacements of the armature and the noise introduced if the 
armature strikes the pole faces. The stiffness required to assure 
stable mid-center adjustment of the armature in the air-gaps intro- 
duces sufficient reactance in most commercial structures of this type 
to place a definite limitation upon the low-frequency response. The 
problem of attaining good response at the higher frequencies is com- 
plicated by the fact that the electrical impedance has a large reactive 
component that often causes appreciable transition losses at these 
frequencies. 

As early as 1877 the suggestion was made that the static force on 
the diaphragm of the magnetic type of receiver could be eliminated 




FIG. 1. Capps' balanced arma- 
ture receiver structure. 






248 



FLANNAGAN, WOLF, AND JONES 



[J. S. M. p. E. 



and its performance improved by making the diaphragm of non- 
magnetic material and attaching to it a driving coil arranged to 
vibrate freely in a transverse magnetic field (Fig. 2). 3 In addition to 
covering the essential features of the moving-coil type of motor 
element, this suggestion also included a flared diaphragm having an 
edge of low stiffness. Obviously, this disclosure incorporated many 
of the structural features that have played an important role in 
modern loud speaker development, for in a well designed motor of 
the moving-coil type the useful driving force is substantially pro- 





FIG. 2. 



Moving-coil receiver introduced 
in 1877. 



portional to the voice current in the coil, even at large amplitudes, 
thus eliminating the double-frequency distortion inherent in the 
magnetic type and making possible the high acoustic outputs required 
of present-day loud speakers without magnetic overloading and dis- 
tortion. Furthermore, the elimination of the steady force developed 
by the polarizing field, and the associated problems of stable air- 
gap adjustment, make it possible to reduce materially the stiffness 
reactance of diaphragm structure, and hence to effect a marked im- 
provement in the response at low frequencies. In addition, the 
electrical impedance varies substantially less over the frequency 
range of interest than it does in the case of the moving-armature 
type of motor element. 



Mar., 1937] 



MODERN THEATER LOUD SPEAKERS 



249 



The problem of transmitting the vibrational energy of a solid body, 
such as a diaphragm, to air is an old one, and in many respects a 
difficult one. It is of interest that the early efforts to solve this 
problem involved the same basic means as are employed today. A 
direct radiator of large area was made by Gray as early as 1877, 




FIG. 3. Direct radiator, invented by Gray. 

Fig. 3. 4 Conical and flared horns were employed in the first telephone 
instruments (Fig. 4). 5 No doubt some of the early horns had rates of 
taper that aided in the transfer of the mechanical energy of the 
diaphragm to the surrounding air. However, horns designed to in- 
corporate an exponential rate of taper appear to be of more recent 
origin. 6 Exponential horns have been investigated theoretically and 
experimentally by Webster and others. 7 
Although the instruments described in the preceding paragraphs 



250 



FLANNAGAN, WOLF, AND JONES [J. S. M. p. E. 



incorporated many of the basic features of present-day loud speaking 
telephones, no source of electrical power large enough to enable them 
to develop sufficient sound to fill a room or auditorium was available. 
Attempts were made to solve this problem with mechanical amplifiers, 
but it was not until the development of the vacuum tube amplifier 
that these principles were applied extensively to uses other than 
telephone receivers, in which case it was necessary to develop only 
enough sound to be audible in the space enclosed between the dia- 
phragm and the ear. With the advent of vacuum tube power ampli- 
fiers, the field of application of the loud speaker was rapidly extended. 

Typical examples of this advance are the 
equipment that was installed along the 
Victory Way in 1918, at the presidential 
conventions of the two major political 
parties in 1920, at the inauguration of 
President Harding in 1921, and at Arling- 
ton on Armistice Day of the same year. 8 
The motor elements of the loud speakers 
during this period were for the most part 
of the balanced-armature type. Several 
different types of horns were employed. 
Included among these were straight horns 
having an exponential taper. In certain 
cases the horns were folded to conserve 




FIG. 4. Early telephone 
employing flared horn. 



space. 

With this background of experience in 
loud speaker design and application to 
draw upon, the developments that made 
possible the high-grade acoustical reproduction that accompanied the 
first showing of sound pictures at the Warner Theater in New York 
in August, 1926, came in rapid succession. It is with these and 
subsequent developments in theater loud speakers that this paper 
is primarily concerned. 

The advances that have been made in the performance of present- 
day loud speakers have been principally along five lines, namely : 

(1) The increase in the efficiency of the loud speaker as an electroacoustical 
transducer, which has resulted not only in outstanding reductions in the power 
losses of the loud speaker itself, but also has materially reduced the amplifier 
capacity required for a given sound intensity in the theater. 

(2) The increase in power-carrying capacity, which has made possible loudness 



Mar., 1937] 



MODERN THEATER LOUD SPEAKERS 



251 



ranges that were hitherto impracticable and which not only enhanced the dra- 
matic possibilities of sound reproduction in theaters but opened new opportunities 
for improvement in recording. 

(5) The extension of the frequency range, both at the high and low ends, 
which has resulted in outstanding improvements in the quality and naturalness 
of reproduction. 

(4) The substantial reduction in non-linear distortion, which has materially 
increased the fidelity of the reproduced sound. 

(5) The improvements in the directive properties of loud speaker units, 
which have aided greatly in attaining uniform results throughout the theater. 

In discussing these advances it is convenient to divide the loud 
speakers into two classes; namely, those in which the entire fre- 
quency range is transmitted by one unit, and those in which the elec- 




FIG. 5. 



Exponential horn equipped with moving-coil 
receiver. 



trical output from the amplifier is distributed between two or more 
units by means of dividing networks. 

Single-Unit Loud Speakers. The loud speakers employed in the 
first showing of sound pictures at the Warner Theater are typical 
examples of the single-unit type. In this installation it was possible 
to obtain for the first time efficiencies as high as 25 per cent over a 
wide frequency range, and to operate at much higher output levels 
than had previously been feasible. It was also possible to attain this 
increase of efficiency and higher output level with a substantial re- 
duction in non-linear distortion. 

The loud speakers employed in this installation were of the horn 
type, and consisted of a moving-coil receiver unit and an exponential 
horn so designed as to conserve the space behind the motion picture 
screen (Fig. 5). Multiple throats permitting the use of either two 
or four units on one horn were also provided. The important struc- 
tural features of the receiver unit are shown in Fig. 6. Although 
the unit has been described in detail in an earlier publication 9 several 



252 



FLANNAGAN, WOLF, AND JONES 



[J. S. M. p. E. 



of the more important structural features will be reviewed here inas- 
much as they represent advances in loud speaker construction typical 
of the recent trends in design. 

It is convenient for purposes of analysis to look upon the loud 
speaker as an electroacoustical transducer drawing power from an 
electrical circuit and delivering it to an acoustical load. The elec- 
trical and acoustical terminal conditions are therefore fixed within 




FIG. 6. Sectional diagram showing the dia- 
phragm, air chamber, and throat construction of 
the 555 type receiver. 

rather narrow limits by the electrical impedance of the associated 
amplifier and the acoustical impedance of the space in which the loud 
speaker is to operate. Achieving high conversion efficiency over a 
wide frequency range is therefore primarily a matter of designing 
the component parts of the structure so as to minimize internal losses 
within this range and, so far as possible, to eliminate transition losses 
due to impedance unbalances. Many of the problems involved in 
obtaining an efficient flow of power from an electrical source to an 
acoustical load have been discussed in previous publications, 10 and 
hence need not be taken up in detail here. However, high magnetic 



Mar., 1937] 



MODERN THEATER LOUD SPEAKERS 



253 



efficiency, low effective vibrating mass and stiffness, and efficient 
radiation play such important parts in accomplishing this result, 
that the means by which this objective was attained in the moving- 
coil unit and horn previously mentioned will be discussed briefly. 
Owing to the limitations imposed by available magnetic materials, 
the permissible size and weight of the magnetic structure, cost, etc., 
the principal problem involved in loud speaker design usually is that 




FIG. 7. High-frequency loud speaker used in 3-way 
system; including sectional diagram showing the dia- 
phragm, air chamber, and horn construction. 

of approaching the ideal as closely as these limitations will permit. 
The design of the magnetic structure of the loud speaker in question 
is, however, not only such that it meets the requirements imposed 
by commercial production, but provides a high force-factor, which 
contributes materially to the high overall efficiency that has been at- 
tained. 

In structures of the horn type it is usually necessary to couple 
the diaphragm to the throat of the horn by means of an air chamber, 
in order to avoid large transition losses. Most loud speakers having 
the conventional form of coupling chamber fail to reproduce the 



254 



FLANNAGAN, WOLF, AND JONES 



[J. S. M. P. E. 



higher frequencies at the same intensity as the lower ones, due to 
the fact that changes in pressure developed at different points on the 
surface of the diaphragm do not reach the throat of the horn in 
proper phase. This effect has been reduced in the design in question 
by adopting an annular throat opening so located between the 
center and the periphery of the diaphragm that disturbances originat- 




FIG. 8. Typical 3-way loud speaker installation. 

ing at the inner and outer portions arrive approximately in phase 
even up to comparatively high frequencies. With this type of 
construction it is possible to use a fairly large diaphragm and deliver 
large amounts of power without an appreciable sacrifice in efficiency 
either at the high or the low frequencies. 

By using thin duralumin sheet and forming it into two reentrant 
spherical surfaces, it was possible to produce a diaphragm not only 
of low effective mass but one that when driven at a point near its 
edge, vibrated substantially as a piston over a wider frequency 



Mar., 1937] 



MODERN THEATER LOUD SPEAKERS 



255 



range than was previously possible. Both these features reacted 
upon the response at the higher frequencies and aided materially in 
improving the efficiency of conversion at these frequencies. 
The driving coil was made of aluminum ribbon wound on edge 




FIG. 9. Experimental 2-way horn system. 

and held together with a thin layer of varnish which served also as 
insulation between the turns. This form of construction provided a 
compact coil having a high ratio of active conductor to total space 
occupied in the air-gap, and permitted operation at higher tem- 
peratures than had previously been tolerable. Moreover, it aided 
materially in obtaining high magnetic coupling, and, because of its 



256 FLANNAGAN, WOLF, AND JONES [J. S. M. p. E. 

small mass, was a contributing factor in improving the efficiency of 
the structure at the higher frequencies. 

As was pointed out in the discussion of various types of motor 
elements, one of the outstanding advantages of the moving-coil type 
is the low stiffness obtainable. Two features of this receiver unit 
aided in attaining this result: namely, a low-stiffness diaphragm 
edge and a hollow center pole-piece which reduced the stiffness of 
the cavity between the center pole-piece and the diaphragm. Tube 
resonance in the pole-piece was prevented by a filling of sound ab- 
sorbing material. 

In addition to having a well designed motor element, it is essential 
that the radiator be an efficient one. In other words, it should be so 
designed as to represent a constant load on the diaphragm throughout 
the frequency range it is desired to transmit. The impedances of 
horns approach constant resistance as the frequency rises, but this 
condition is attained most rapidly when the cross-sectional area 
varies exponentially with distance along the axis. A unique feature 
of such a horn is that its transmission range is inherently limited at 
the low-frequency end. An infinite horn of this type would suppress 
the frequencies below this limiting value and transmit freely the 
frequencies above. The closeness with which a finite horn approxi- 
mates these characteristics depends upon the efficiency with which 
the mouth radiates energy to the surrounding air. The mouth of 
the horn should, therefore, be of such dimensions that it will radiate 
freely waves of the lowest frequency in the desired range, and the 
conduit portion should be capable of freely transmitting all frequencies 
that the mouth will radiate. It is this property of the rate of taper, 
that all frequencies in the range of efficient radiation of the mouth 
opening are freely transmitted, that differentiates the horns of recent 
design from the earlier types mentioned in the first part of this paper. 10 

A large number of these single-unit loud speakers have been in- 
stalled in theaters since their initial introduction in 1926 and have 
given excellent service. Many of these instruments have never been 
replaced by units of later design and are still in daily use. 

Multiple-Unit Loud Speakers. While it has proved feasible as a 
result of careful design to extend markedly the frequency and volume 
ranges of the loud speakers of the single-unit type, the fact that the 
entire frequency range must be covered by one unit imposes important 
design limitations. In general, a diaphragm designed for efficient 
radiation of low frequencies has too high a mass to be best suited 



Mar., 1937] MODERN THEATER LOUD SPEAKERS 257 

for use at high frequencies. In addition, the problem of radiation at 
the high frequencies is further complicated by the tendency of a 
diaphragm large enough to meet the low-frequency requirements to 
change its mode of vibration and no longer act as a piston as the 
frequency is increased. Furthermore, a diaphragm designed to 
radiate the high frequencies efficiently is not best suited mechanically 
to develop the large amounts of acoustical power required at the 
lower frequencies without serious distortion and possible mechanical 
failure. 

Considerations such as these led first to the development of a 
moving-coil loud speaker of the horn type having vibrational charac- 
teristics permitting efficient radiation at the higher audible fre- 
quencies. 11 A model of this speaker is shown in Fig. 7, with a sec- 
tional sketch showing the internal construction. The diaphragm 
was made of duralumin drawn into a hemispherical dome at the cen- 
ter. A voice-coil of aluminum ribbon was attached at the edge of 
the dome portion. The horn was tapered exponentially and termina- 
ted in an annular throat, to minimize high-frequency interference 
in the cavity in the front of the diaphragm. The mechanical and 
acoustical constants of the various portions of the structure were so 
chosen as to result in efficient radiation at frequencies as high as 
12,000 cps. with substantially uniform response at all frequencies 
between 3000 and 12,000 cps. This speaker was used as an adjunct 
to speakers of the single-unit type, and resulted both in a substantial 
increase in the frequency range and in a marked improvement in the 
distribution of the high-frequency sounds. 

Three- Way System. The design of loud speakers especially suitable 
to radiate the low frequencies was next undertaken with such suc- 
cess that three-band systems became especially attractive and 
eventually led to the commercialization by Electrical Research 
Products, Inc., of the wide-range system. 12 In this system the output 
from the amplifier is distributed between three loud speakers by 
means of dividing networks. The first unit covers the range up to 
300 cps., the second from 300 to 3000, and the third from 3000 cps. 
up (Fig. 8). 

In this system the horn-type moving-coil loud speaker described 
in connection with the single-unit system was retained as the mid- 
range unit. This not only made it possible to increase materially 
the maximum output level of this unit without adding to the dis- 
tortion, but also had the practical advantage that it made the maxi- 



258 FLANNAGAN, WOLF, AND JONES lj. S. M. P. E. 

mum use of existing equipment when extending the range of a system 
already in operation. The high-frequency unit was the one described 
above, and was used both singly and in multiple for increasing the 
power and the capacity. 

The low-frequency unit originally consisted of a short horn having 
an exponential taper and three dynamic loud speakers having coni- 
cally shaped paper diaphragms. It was found, however, that better 
response at the lowest frequencies could be attained by using a large, 
flat baffle. This baffle was of rigid construction, and was so braced 
that the resonance frequencies of the various areas were not alike. 
This construction proved to have a decided advantage in that the 
low frequencies were reproduced with noticeably less distortion. 

It was found by experience that the positioning of the units, 
particularly the high-frequency units, had an important reaction upon 
the quality obtained with the system. In setting up such a system 
it was necessary to adjust the position of the high-frequency unit 
until a location giving the most pleasing quality was found. 12 This 
process is often referred to as "phasing," and when properly carried 
out resulted in a marked reduction in distortion, especially in con- 
nection with the reproduction of sounds of the impact type. 

It has been stated that poor directional characteristics 13 constitute 
one of the principal limitations of the present theater installations. 
A solution of this problem was found during the development of the 
loud speakers used in connection with the reproduction of speech 
and music in auditory perspective. One of these loud speakers is 
shown in Fig. 9, and consists of a large-diaphragm moving-coil unit 
with a folded horn for reproducing the frequencies up to approxi- 
mately 300 cps., and a multi-throat moving-coil unit with a cellular 
type of horn for reproducing the frequencies above 300 cps. 13 The 
first public demonstration of a system employing these loud speakers 
was given in April, 1933, when the music of the Philadelphia Orches- 
tra, playing in the Academy of Music, Philadelphia, was reproduced 
stereophonically in Constitution Hall, Washington, D. C. 

It is a well known fact that the sound from the conventional type 
of tapered horn is projected through a relatively large area at the 
lower frequencies, but that as the frequency is increased the distribu- 
tion becomes less and less uniform until at the frequencies where 
the wavelength is small as compared with the mouth opening, the 
output of the horn is concentrated in a narrow beam along the axis. 
Concentration of the high frequencies in theater reproduction re- 



Mar., 1937] 



MODERN THEATER LOUD SPEAKERS 



259 



suits in a situation in which if a proper balance between high and 
low frequencies is attained on the axis of the horn, the sound in the 
regions to either side tends to be deficient in high frequencies and 
is characterized as "boomy." Conversely, if the high and low 
frequencies are balanced at a location off the beam, an excess of high 
frequencies on the axis of the horn results. Obviously, this situation 
would be much improved if the radiation from the horn were not 
spherical but were confined to a solid angle such that the undesirable 




FIG. 10. Sectional diagram showing the 
diaphragm, air chamber, and throat construc- 
tion of experimental high-frequency unit. 



effects of reflection from side walls, ceiling, etc., were kept to a mini- 
mum. 

This was accomplished in the auditory perspective system by 
adopting a multi-cellular type of horn composed of a number of 
separate channels, each having substantially an exponential rate of 
taper. The narrow ends of the channels were brought together with 
their axes parallel and were terminated in a single tube which, in 
turn, was connected to the receiver unit. Sound from the unit 
travels down this tube and divides among the channels. For fre- 
quencies of which the wavelengths are large compared with the 
mouth opening, a substantially spherical wave results. As the 
frequency is increased the sound tends to concentrate more and more 
upon the axis of each channel. This is permissible, provided the 



260 



FLANNAGAN, WOLF, AND JONES 






[J. S. M. P. E. 



construction of the horn is such that the sound coming from the 
individual channels is normal to a spherical wave-front. 

The receiver unit employed with the multi-cellular horn is shown 
in Fig. 10. As in the case of the motor elements of the single-unit 
loud speaker and the high-frequency unit of the three-way system 
previously described, this unit is of the moving-coil type, and has 
a formed diaphragm of thin duralumin and a voice-coil of aluminum 





FIG. 11. Diphonic loud speaker system. 

ribbon. The throat is also annular but three openings instead of 
one are provided in order to increase the frequency range. 

Diphonic Loud Speaker. While the wide range system represented 
a marked improvement from the standpoints of frequency and volume 
ranges, it was not fully adequate in several respects. The distribu- 
tion of sound in the seating area was not materially improved, owing 
to the fact that the same mid-range loud speakers were employed. 
Furthermore, there was still sufficient distortion due to transients, 
frequency modulation, and nonlinear effects under certain conditions 
to be disturbing to the critical listener. 



Mar., 1937] MODERN THEATER LOUD SPEAKERS 261 

These deficiencies are partly, but by no means entirely due to the 
loud speakers. Therefore, improvement in the other parts of the 
sound system was necessary in order to effect a distinct advance in 
reproduced sound. Advances in the component parts have been 
made, and the combination of equipment embodying these advances 
is now know as the Mirrophonic System. The loud speaker of this 
system has been termed "Diphonic," and is shown in Fig. 11. 

The Diphonic loud speaker is a two-way loud speaker in which the 
output from the amplifier is distributed between the two units by 
means of a dividing network that introduces approximately 12 




FIG. 12. Multicellular horn equipped with two high- 
frequency units. 

decibels of attenuation per octave above and below the cross-over 
point of 300 cps. 

The high-frequency unit and horn are shown in Fig. 12. The 
receiver unit is a commercial form of the one used in the Philadelphia- 
Washington demonstration previously mentioned. The physical 
form of the multi-cellular horn has, however, been changed somewhat, 
without deviating from the basic design considerations mentioned 
above. The original horn was driven by one receiver; most of the 
commercial horns are driven by two units. This is accomplished by 
using a F-type throat, and permits the use of the horn in theaters 
requiring more power than can be handled by one driving unit. 
For smaller theaters, requiring less power, a casting is provided that 



262 FLANNAGAN, WOLF, AND JONES [J. S. M. p. E. 

permits the use of a single unit. It will be noted that the space 
between the horn cells is open in the commercial unit but closed in 
the experimental unit. In the experimental unit the space between 
the horn cells was acoustically deadened by cotton waste packing, 
which represented about 30 per cent of the weight of the combined 
horn and unit. The commercial horn eliminated the waste packing 
by including the deadening medium in the material of the individual 
cell. These changes are the results of experimental data obtained 
in a long series of tests and have accomplished a simplification of the 
design, reduction in weight, and further improvement in the dis- 
tribution characteristics of the horn. The high-frequency unit now 
employed in the Diphonic loud speaker gives a substantially uniform 
distribution of sound with respect to both volume and frequency over 
an angle of 90 degrees, and has been used satisfactorily over consider- 
ably wider angles. Where a critical listener could detect a certain 
amount of beam effect even in the best wide-range installation, this 
effect is practically unnoticeable in the Mirrophonic system, with the 
result that "boominess" is avoided and the intelligibility as well as 
the naturalness of dialog is greatly improved. 

The low-frequency unit also shows marked improvement over the 
low-frequency speakers previously used commercially. As with the 
case of the high-frequency unit, this device is based upon extensive 
experimental data obtained in a variety of tests. Here the problem 
is one of maintaining minimum physical depth with high translation 
efficiency. To accomplish this it is desirable to utilize a large driving 
area. Consequently, a system employing four dynamic speakers of 
the cone type was adopted. Each of these cone units is about 16 
inches in diameter and is coupled to the base of a shallow cavity 
surrounded by a rigid baffle. The parts forming the cavity and 
baffle are so rugged that spurious sound radiation due to mechanical 
vibration of the baffle is prevented. Conventional data were used to 
determine the general cavity dimensions, but preliminary models 
indicated conclusively that considerable departure from theoretical 
views was necessary in order to produce a coupling device free from 
resonance effects causing a prolongation of certain tones beyond their 
natural duration, thus masking succeeding tones and impairing the 
quality. 

During the development of the Diphonic loud speaker the con- 
struction of test models was guided by listening tests as well as acous- 
tic measurements. Subsequently to the attainment of satisfactory 



Mar., 1937] MODERN THEATER LOUD SPEAKERS 263 

response from the standpoint of frequency characteristic and effi- 
ciency, a tendency toward "beaminess"of the high-frequency radiation 
of the low-frequency unit was observed. This was corrected by the 
addition of vanes that effectively distributed these sounds over a 
horizontal angle comparable to that covered by the high-frequency 
horn previously described. 

In addition to the wider frequency range, reduced distortion and 
better directional properties previously mentioned, the improved 
efficiency of both units has definitely increased the volume range. 

The Diphonic loud speaker requires a minimum of space on the 
stage; can be quickly dismantled if required; and the technic followed 
in its installation is considerably simpler than that required by some 
of the older equipment. 

REFERENCES 

1 WATSON, T. A.: U. S. Pat. No. 266,567; filed April 17, 1880, issued October 

24, 1882. 

I CAPPS, F. L.: U. S. Pat. No. 441,396; filed April 7, 1890, issued November 

25, 1890. 

SIEMENS & HALSKE: German Pat. No. 2355; December 14, 1877. SIE- 
MENS: Brit. Pat. No. 4685; February 1, 1878. CUTTRISS AND REDDING: U. S. 
Pat. No. 242,816; June 14, 1881. 

4 PRESCOTT, G. B.: "Bell's Electric Speaking Telephone," D. Appleton & Co. 
(New York), 1884, p. 458. 

*Ibid., p. 441. 

TOURNAYRE AND TouRNAYRE: French Pat. No. 463,660; submitted October 
15, 1913, published March 2, 1914. 

7 WEBSTER, A. G.: "Acoustical Impedance and Theory of Horns and Phono- 
graph," Proc. Nat. Acad. Sci., S (1919), No. 6, p. 275. 

8 GREEN, I. W., AND MAXFIELD, J. P.: "Public Address Systems," Bell Syst. 
Tech. J., 2 (April, 1923), No. 2, p. 113. 

9 WENTE, E. C., AND THURAS, A. L.: "High-Efficiency Receiver of Large 
Power Capacity," Bell Syst. Tech. J., 7 (Jan., 1928), No. 1, p. 140. 

10 MAXFIELD, J. P., AND HARRISON, H. C.: "Methods of High-Quality Record- 
ing and Reproducing Music and Speech Based on Telephone Research," Bell 
Syst. Tech. J., 3 (July, 1926), No. 3, p. 493. HARRISON, H. C.: U. S. Pat. No. 
1,730,425; October 8, 1929. 

II BOSTWICK, L. G. : "An Efficient Loud Speaker at the Higher Audible Fre- 
quencies," /. Acoust. Soc. Amer., 2 (Oct., 1930), No. 2, p. 242. 

11 MAXFIELD, J. P., AND FLANNAGAN, C. : "Wide-Range Reproduction in 
Theaters," /. Soc. Mot. Pict. Eng., XXVI (Jan., 1936), No. 1. p. 67. 

11 WENTE, E. C., AND THURAS, A. L.: "Symposium on Wire Transmission of 
Symphonic Music and Its Reproduction in Auditory Perspective Loud Speakers 
and Microphones," Bell Syst. Tech. J., 13 (April, 1934), No. 2, p. 259. 



264 FLANNAGAN, WOLF, AND JONES 

DISCUSSION 



MR. KELLOGG: In practically all these loud speaking systems in which one 
speaker handles the low frequencies and the other the high, there is a difference 
in the time at which the sound from the two sources reaches the listener, corre- 
sponding to four to six or eight feet of travel, since the high-frequency unit is 
placed nearer the front of the horn of the low-frequency speaker. Has any 
definite conclusion been reached as to the effect of the difference in the time of 
arrival of the two components? 

MR. JONES: The effect is referred to in theater reproduction as "phasing." 
It plays a very important part, particularly with the impact type of sound. 
It is a design feature that I did not have time to discuss; however, it is important 
in arranging a loud speaker system, particularly one involving a large number of 
units of different horn lengths, that the diaphragms be so located relatively to 
the face of the unit as a whole as to correct the effect that you have in mind. 

This is usually done in the case of multi-unit systems by adjustment in the 
theater. It is possible, however, if the units have horn lengths essentially 
equal, to take care of the major portion of the effect in the initial assembly. 

MR. COOK: Loud speakers are generally mechanical circuits having series- 
connected elements at their fundamental resonance frequency. This frequency 
is usually either below or near the lower limit of their frequency response. In 
addition, these devices are seldom critically damped at this resonance frequency . 
In such circuits any applied impulse would create an oscillatory transient that 
would not die out immediately upon removing the applied force. 

There have been many examples of loud speakers, the transients of which 
have a duration of as much as half a second. Practically any musical composi- 
tion would have notes changing from one tone to another in time intervals of 
this order, at the same time being preceded by transients having relatively steep 
wavefronts. 

It would seem probable that completely faithful reproduction would be im- 
possible under such conditions, for while reproducing a desired tone, another 
that had occurred at some previous time might still be present. Have you any 
information as to how serious that has been, and whether any marked improve- 
ments in that regard have been made or are likely in the near future? 

MR. JONES: Speakers such as the ones I described represent an improvement 
in that respect. I feel quite confident that in the future we may expect greater 
improvements from the standpoint of transient characteristics. 

MR. CRABTREE: What is the power capacity of the units especially the 
large units? 

MR. JONES: The objective in present-day reproducing systems for theaters 
is to develop about 25 or 30 watts of sound power. Measurements made on 
orchestral instruments show that maximum power occurs in the region of 250 
to 500 cps., and is of the order of ten watts. A system that has an output ca- 
pacity of 25 to 30 watts is capable of delivering from two to three times the 
maximum power of a sizeable orchestra. 

With good conversion efficiency, that is, of the order of 50 per cent, this means 
that the electrical input levels can be very materially reduced in obtaining the 25 
to 30 watts, and hence the amplifier problem is very much simplified. There is 
a fair factor of safety over that. I should say that it is not unreasonable to 
expect 100 per cent overload to be carried for a very short time. 






RESEARCH COUNCIL SPECIFICATIONS FOR A 
STANDARD SYNCHRONIZING SYSTEM FOR CAMERAS* 



Summary. / July, 1936, the project of adopting a standard synchronizing 
system for cameras was suggested to the Research Council of the Academy of Motion 
Picture Arts & Sciences for the purpose of eliminating confusion on stages resulting 
from the use of clap-sticks and other visual-auditory synchronizing devices, to derive 
from the system the advantages of adopting uniform methods in all studios, and to 
achieve such uniformity among cameras as will allow using the cameras in all studios 
without altering the synchronizing equipment. Specifications of the standard syn- 
chronizing system are given for cameras of several makes. 

SUBJECT 

These specifications describe the combination and application of 
existing and/or modified commercial equipment to a system to syn- 
chronize the camera with the sound recording equipment by fogging 
the film inside the camera after the equipment is in motion and has 
reached operating speed. 

These specifications are designed to cover the simple fundamentals 
of a synchronizing system, and may be elaborated upon in any way 
provided the fundamental dimensions and units of equipment are 
retained within the system. 

The Research Council drawing entitled Units of the Research 
Council Standard Synchronizing System for Cameras (Fig. 1) shall 
be considered a part of these specifications. 

The drawings entitled "Application of the Standard Synchronizing 
System for Cameras to the Mitchell NC Camera" (Fig. 1), "Applica- 
tion of the Standard Synchronizing System for Cameras to the Stand- 
ard Mitchell Camera" (Fig. 2), and "Application of the Standard 
Synchronizing System for Cameras to the DeBrie Super- Parvo Cam- 
era" (Fig. 3) shall not be considered a part of these specifications but 
are presented only to assist in the application of the standard to 
actual studio practice. 



* Reprinted from the Technical Bulletin of the Academy of Motion Picture 
Arts & Sciences, December 31, 1936; effective January 1, 1937. 

265 



266 



RESEARCH COUNCIL SPECIFICATIONS [J. S. M. P. E. 

STANDARD UNITS 



The individual items of which this standard system consists have 
been so chosen that either an argon vapor lamp and socket or a 
filament-type switchboard lamp and socket may be used interchange- 



SOCKET AND PLUG 
6 E. "1346 4 1347 or equ.vtlcnt 




= | or 6 v P C 



WIRING DIAGRAM 



SWITCHBOARD LAMP 
* SOCKET 




-7J xSV-P 

-T ( ) dependent upon 

_i_ : X <4 ^ type of camera 



Plug, Adapter, Lamp 4 Cap assem 
ARGON LAMP ASSEMBLY 



FIG. 1. Units of the Research Council standard synchronizing system for 
cameras (October 26, 1936). 

ably in any camera and/or any camera equipped with this standard 
synchronizing system may be connected into any studio power 
supply system with proper electrical characteristics, to provide a 
standard synchronizing mark upon the film. 



Mar., 1937] RESEARCH COUNCIL SPECIFICATIONS 



267 

In addition, the individual items included in this system have been 
chosen of dimensions and type which are inexpensive and commer- 
cially available (in some cases with slight modifications). 




FIG. 2. Application of the Research Council standard synchronizing system 
to the Mitchell NC Camera (October 26, 1936). 

THE SYSTEM 

The Standard Synchronizing System for Cameras shall consist of: 
1) a standard plug of dimensions shown in Fig. 1 and into which 
rither an argon vapor lamp or a filament-type switchboard lamp 
vith an appropriate socket will fit interchangeably, mounted inside 
he camera in such a position as to throw a beam of light upon the 



268 



RESEARCH COUNCIL SPECIFICATIONS [J. S. M. p. E. 



film, positioned at a fixed distance from the center of the aperture; 
(2) external resistors of sufficient size and connected as shown in 
Fig. 1, to provide the proper current for either the argon vapor or 




x ~-Carnr in Viewing Position 




FIG. 3. 



Application of the Research Council standard synchronizing system 
to the Mitchell standard camera (October 26, 1936). 



filament- type switchboard lamp; (3) a standard General Electric 
1346 (or equivalent) plug and an associated General Electric 1347 
(or equivalent) socket through which the synchronizing lamps may 
be connected to either a 110- volt a-c. (for the argon vapor lamp) or 
a 6-volt d-c. (for the filament-type lamp) current supply. 



Mar., 1937] RESEARCH COUNCIL SPECIFICATIONS 

DIMENSIONS 



269 



The socket, into which an argon vapor or filament-type switch- 
board lamp will fit interchangeably shall be of fundamental dimen- 
sions shown in Fig. 1, and shall be so located that there will be a 



e^f. * +,\ 



..Jr 




FIG. 4. 



Application of the Research Council standard synchronizing system 
to the DeBrie super-parvo camera (December 17, 1936). 



clearance of not more than 3 /ie or less than Vie inch between the end 
of the lamp-and-socket assembly and the film. 

The outside plug (General Electric No. 1346, or equivalent) shall 
have a maximum length of iy 4 inches, width of 7 / 8 inch, and a height 
of 17 /32 inch, with plugging elements Vw by Y 4 inch in vertical di- 
mensions, located 3 / 8 inch from center to center. 

The outside socket (General Electric No. 1347, or equivalent) 
| shall have a maximum length of V/ 2 inches, width of 7 / 8 inch, and 
height of 17 /32 inch, with sockets of suitable dimensions into which 
the lugs of the above plug will fit. 



270 RESEARCH COUNCIL SPECIFICATIONS 

MOUNTING 

The synchronizing lamp socket shall be so mounted within the 
camera that a beam of light will be projected upon the film covering 
the sound-track area and extending into the picture area. 

The synchronizing lamp socket shall be so mounted within the 
camera that a beam of light will be projected upon the film covering 
the sound-track area and extending into the picture area. 

The lamp and socket assembly shall be so positioned that the 
synchronizing mark shall be located at a point exactly II 1 / 2 frames 
from the center of the camera aperture in the Mitchell NC and the 
Mitchell Standard cameras, and at a point exactly 17 frames from 
the center of the camera aperture in the DeBrie Super- Parvo cameras. 

Note. Although not included as part of these specifications, it 
is suggested that an outline of film loop be etched upon the camera 
wall to indicate the exact length of film with which the camera is to 
be threaded in order to position the synchronizing mark accurately. 

To conduct the studies in the project described above, S. J. 
Twining was appointed by the Academy Sound Recording Committee 
to investigate the current status of camera synchronizing methods, 
both with reference to studio-owned as well as commercially owned 
cameras available by rental to the producing companies. 

A consulting committee, consisting in general of the studio camera, 
laboratory, and cine-technical department heads, was appointed to 
determine the various units of the system that would best meet the 
various requirements of the industry. This committee consisted of: 

J. ARNOLD F. FLACK J. M. NICKOLAUS 

F. S. CAMPBELL F. GAGE E. OSTER 

G. CRANE G. LABUE W. G. ROBINSON 
W. EGLINTON M. LESHING W. RUDOLPH 

H. ENSIGN C. LINDBLOOM A. TONDREAU 

G. FISCHER E. B. McGREAL R. WILKINSON 

A Sub-Committee consisting of W. C. Miller, H. G. Tasker, S. J. 
Twining, and Gordon S. Mitchell, Manager of the Research Council, 
represented the Sound Recording Committee in the discussions of 
the projects with the camera and laboratory groups. 

The Chairman of the Research Council is William Koenig; Sound 
Recording Committee, E. H. Hansen; Sub-Committee on Standardi- 
zation of Synchronizing Systems for Cameras, S. J. Twining. 



TRANSMISSION OF SOUND AND VIBRATION IN 
BUILDINGS* 



ERWIN MEYER** 

Summary. A description of studies made at the Institut fur Schwingungfor- 
schung, Berlin, in the transmission of sound and vibration in buildings, with par- 
ticular reference to the analogies between the properties of sound insulating structures 
and electrical networks. Some of the problems treated deal with the air-borne trans- 
mission of sound through multiple walls; tbe propagation of sound in building 
materials, and the physical properties of insulating materials for structure-borne 
sound; and electrical apparatus for measuring vibrations in buildings. 

As is known, a distinction is made between air-borne and struc- 
ture-borne sound. Transitions between these two classes of trans- 
mission occur very often and cause many difficulties in sound and 
vibration insulation ; on the other hand, the various mutual relations 
between them and the analogies to corresponding electric questions 
are very interesting. There are so many problems of this sort that 
only a few especially treated in this Institute will be selected: air- 
borne transmission through multiple walls; sound propagation in 
building materials, and physical properties of insulating materials 
for structure-borne sound; and electrical apparatus for measuring 
vibrations in buildings. 

MULTIPLE WALLS 

From the measurements of Berger in 1911 it is known that the 
sound insulation of a single wall depends almost entirely upon its 
mass. This is quite right, if the lowest natural period of the wall is 
several octaves under the used frequency range and if the damping 
is not too small. This law can be theoretically demonstrated, and is 
also confirmed by experiments. For the usual building materials and 
partition walls of ordinary dimensions the previous assumptions are 
not completely fulfilled; thicker and heavier walls especially have 
higher and less damped natural periods ; therefore we do not get the 



* Received October 10, 1936. 
** Institut fur Schwingungforschung, Berlin. 



271 



272 



E. MEYER 



[J. S. M. P. E. 



theoretically expected weight- insulation curve of single walls. The 
weight curve has a very important meaning, but is only empirical. For 
the theory of multiple walls we assume that a section of the multiple 
wall is physically like a mass. Fig.l shows the electrical analogue of 




M 





motor 



tight source 



FIG. 1. (Upper.) Multiple wall (scheme and electrical 
analogy). 

FIG. 2. (Lower.) Automatic apparatus for recording 
sound transmission. 

the multiple wall. The mechanical-acoustical behavior resembles 
that of a low-pass filter. The cut-off frequency is 



ml 

where m is the mass (g/cm 2 ) of the single wall, / the thickness (cm) 
of the air space, p the air density (g/cm 3 ), and c the sound velocity 
(cm/sec). The arrangement for measuring the sound transmission 



Mar., 1937] 



SOUND AND VIBRATION IN BUILDINGS 



273 



of such a wall is shown in Fig. 2. The multiple wall under test sepa- 
rates two empty rooms having large reverberation times, so that we 
ihave in each room a diffuse sound field. First, the sound density in 



Sound 
insulation 



fit 2 3 S 6 7 8 1000 Z 3 

FIG. 3. Sound insulation curve: sound pressure (a) in 
transmitter room; (>) receiver room. 

-he loud speaker room is recorded with a logarithmic valve- volt- 
meter; then the density in the second room. The entire procedure 
then repeated. An original photographic record of a wall is shown 
in Fig. 3. 

sound insulation 



















" 


SW,W 






















/ 

/ 


? 


^ 


/^ 























i 


f 


/ 


r 






















rf 


V 


/ 






















^== 


== 


V 




















































SO 100 I 3 U 5 S 8 1008 2 J 9 S ( 1 It 



J 



199 2 J 9 i t I ItM I J * S f I 10009 

FIG. 4. Multiple wooden wall with space 
damping; electrical low-pass filter. 

A series of experiments done by this method showed that triple, 
quadruple, quintuple, etc., walls have cut-off frequencies almost in 
igreement w'th the formula above. Still another effect known from 



274 



E. MEYER 



[J. S. M. P. E. 



the principles of propagation of electric currents in filters was present: 
the phase velocity of a tone is less when its frequency approaches the 
cut-off frequency. An impulse is thus dispersed into its frequency 
components, and it makes a great impression upon an observer when 
hand clapping in front of the wall is heard as the sound oo-e behind 




0- 

20- 



velocity 



\ 



r\ 

& Vi 



bending wave 



\ 3 




/ 12 3t 

** 



FIG. 5. (Upper) Concrete rod with driving apparatus 
(a) and pick-up (V) ; (lower) distribution of amplitudes. 

the wall ; the low frequencies arrive first, as can be demonstrated by 
an oscillogram. But all measurements made on multiple walls showed 
one difference between the corresponding mechanical and the elec- 
trical low-pass filters. An electrical filter has a large transmission 
loss beyond the cut-off frequency; the corresponding mechanical 
filter shows only a slow increase of attenuation, the reason being 
that the .a.ir cushion between two sectional walls does not represent 



Mar., 1937] 



SOUND AND VIBRATION IN BUILDINGS 



275 



a real stiffness. The cushion is, in fact, quasi-stationary in the thick- 
ness dimension; however, the dimensions parallel to the wall are 
greater than or at least comparable to the wavelength of the tones. 



t) ***** 




^ concrete 

iron scaffolding) houses 

brick. ) 

* iron scaffolding 
with structure -borne 
sound insulation 



'123*56793 

floors underneath sound source 



FIG. 6. 
FIG. 7. 



Decrements of concrete rod. 



(Upper.} 

(Lower.) Propagation of structure-borne sound 
in buildings. 



The diffuse incidence of sound creates a pressure gradient parallel to the 

walls, and in order to avoid resonances in this direction the air spaces 
' between the walls must be damped by inserting porous absorption 

materials. It is sufficient to fix these materials at the boundaries. 
1 For a multiple- wall construction of this kind (wooden plates, m = 0.2 

g/cm 2 , / = 3 cm, o = 480 cps.) the results are shown in Fig. 4. 



276 



E. MEYER 



[J. S. M. P. E. 



On the right side of the figure is drawn for comparison the attenuation 
curve of a low-pass filter the exact electrical analogy of the wall. 
The quintuple wall corresponds to a network having four sections. 
The curves are very similar except at the higher insulation values, 





FIG. 8. Mechanical impedance meter; section and electrical circuit. 

where they are lowered in the mechanical case by secondary influ- 
ences. If side-damping is applied to the air spaces, making the thick- 
ness equal to half a wavelength, air-space resonances in the thick- 
ness direction are distinctly obtained. If side-damping is not used, 
these resonances can not be found. To avoid them the entire air 



X 



FIG. 9. Reactance (left) and resistance (right) curves 
for cork-rubber-cork combination (ordinates X 10 g/sec. 
mechanical ohms). 

space must be damped, but in the most cases the air space is small in 
comparison with the wavelength. Multiple walls manufactured 
according to these principles with a cut-off frequency lower than 100 
cps. insulate surprisingly well, giving 20 db. more attenuation than 
a single wall of the same weight. The formula given above for the 
cut-off frequency shows that for good insulation it is necessary to 



Mar., 1937] 



SOUND AND VIBRATION IN BUILDINGS 



277 



lower this frequency, which can be done by using heavier sectional 
walls or thicker air spaces. That means that for good insulation 
against air-borne sound the requirements are either weight or plenty of 
space. 

The loosest practically possible coupling between two walls is an 
air space between them. In most cases there must be additional 
solid connections between the walls, which constitute at the same 
time sound bridges. We have studied these questions to son e 
extent, and have found, of course, parallels to a telephone line. A 
short line having continously distributed capacitance and inductance 
can be inductive or capacitive ac- 
cording to its terminating impedance. 
In the case of light walls, heavy and 
elastic sound bridges transmit less 
than do light bridges of less elas- 
ticity. The converse is the fact in 
the case of heavy walls. In every 
case we have found that supports 
with springs and rubber, which arc 
the constructional elements of the 
floating walls, are the best. 



lever 




for vertical component 



nil 



STRUCTURE-BORNE SOUND 

Sound bridges lead directly to the 
question of sound propagation 
through a building as a whole. First 
must be considered the attenuation 
in the building materials themselves, 
and we must distinguish between 
the various types of sound waves 
and transverse or bending waves. 



T 

.};;:J 



For horizontal component 



FIG. 10. 



Electrodynamic 
lion meter. 



vibra- 



e. g., longitudinal, torsional, 
The transmission of these sound 
waves can easily be measured by a resonance method. A bar of the 
material in question (wood, concrete, brick, stone, iron) is driven at 
one end by an electrodynamic system of low mechanical resistance 
(at a, Fig. 5). The velocity amplitude is measured by a gramo- 
phone pick-up at the other end of the rod (b, Fig. 5). In Fig. 5 is 
shown the curve of velocity amplitude along the rod for a transverse 
resonance (of the fourth order, to show the characteristic displacement 
of the ends) and for a longitudinal vibration (of the second order). 
To get the attenuation of the waves, one has only to measure the 



278 E. MEYER [J. S. M. P. E. 

resonance curves and to determine the decrement 6. Assuming low 
damping, the theory shows that the attenuation of the waves in the 
distance \/2ir is 6/2ir or 0/4T, respecitvely, for longitudinal and tor- 
sional or transverse waves, respectively. In Fig. 6 are represented 
the decrements for two concrete bars of different lengths, for the dif- 
ferent kinds of waves; the decrements are almost independent of the 
frequency. This leads to the conclusion that the losses are material 
losses due to hysteresis. 

For a technical survey it is better to express the physical data in 
decibels per meter. The data differ, of course, for the different 
pitches and waves, and one can state only the order. To attain an 
attenuation of 1 decibel the following lengths are required: iron, 
25-1000 m.; brick, 8-50 m.; concrete, 5-30; wood, 3-20. No pro- 
spect therefore appears of damping the sound waves by pure material 
losses. Damping can occur only by transition from one building 
structure to another. In every building there are such transitions 
which give a more or less sufficient damping. This is illustrated by 
Fig. 7. In various houses in Berlin, of brick, iron, or concrete struc- 
ture, the sound propagation in the structure itself has been measured. 
In Fig. 7 the abscissas represent the number of floors underneath the 
sound-source, the ordinate the diminution of loudness in decibels. 
The sound-source (trampling machine) was generally placed at the 
highest floor. The ordinary brick building has the greatest trans- 
mission loss (19 decibels per floor), while concrete and iron have only 
8 decibels per floor. By insulating the iron scaffolding with cork 
plates (Curve 6, Fig. 7) the same values are attained as for the brick 
houses. 

To interrupt the sound propagation in structures we use "sound- 
soft" materials: namely, materials having low impedance, such as 
cork, rubber, etc. It is important to know the physical properties of 
these materials. After many experiments an electromechanical 
resistance meter was constructed for this purpose. Two equal pieces 
of the material to be tested, separated by a metal disk (Fig. 8), are 
placed between the poles of two pot-shaped magnets. A coil C\ is 
situated in the circular air-gap of the magnet and fixed to the metal 
disk. A current flowing in this coil exerts a force upon it, and con- 
sequently also upon the test pieces. 

Similarly, on the other side of the metal plate is another coil. The 
voltage induced in it measures the velocity amplitude of the metal 
disk and the test pieces. The test pieces are loaded by inserting 



Mar., 1937] 



SOUND AND VIBRATION IN BUILDINGS 



279 



metal sheets, if an additional static pressure is desired. To prevent 
the current coil from influencing the voltage coil, each is divided into 




Fio. 11. Electrodynamic vibration meter. 

parts wound in opposite sense, half the winding being fixed and 

other half movable. Fig. 8 shows the electrical measuring 

ivice; the induced emf. is compensated by a Larsen compen- 



280 



E. MEYER 



[J. S. M. P. E. 



sator. If Z = r + juM is the impedance of the compensator, we 
obtain for the mechanical impedance of the whole system (metal plate 
and material under test) W = AiA z /Z. The constant AiA 2 is deter- 
mined by means of a known reactance, e. g., by a mass. The me- 
chanical impedance of the system by itself, without test pieces, is 
measured similarly. The frequency range used is 30-700 cps. 

What results are to be expected? For the lower frequencies a 
material such as cork, felt, rubber, etc., is equivalent to a resistance 
connected in series with a capacity. The capacity corresponds to a 
mechanical compliance C, and the electrical resistance to a mechani- 
cal resistance B. The mechanical reactance B is thus equal to l/o>C 
and it remains to find the dependence of the resistance R upon the 

Kef lection for constant 



1,0 



0.5 



1 IS I 96 10 W W 60 80100 

FIG. 12. Response curve of vibration meter. 

frequency. If we assume that a certain fraction a of the (potential) 
input power Bv 2 (v = velocity amplitude) is transformed into the heat 
Rv 2 , it follows that R is proportional to 1/w, because B = 1/coCand 
C itself are constant. These assumptions are, in fact, confirmed by 
experiment. For cork with a rubber layer, Fig. 9 illustrates the de- 
pendence of the reactance and the resistance upon the frequency, the 
two curves referring to different static pressures. Both have a 
slope of 45 degrees; which means, for the reactance curve, that the 
compliance and, therefore, also the elasticity constant, do not change 
with the frequency. The greater the loading the smaller is the com- 
pliance. The resistance curves show that the resistance is inversely 
proportional to the frequency. Corresponding to the dielectric 
phase difference of a condenser, a loss factor can be determined for the 
material, usually termed "absorption factor" (but in another sense, 
as in room acoustics). Generally the ratio of resistance to impedance 



Mar., 1937] 



SOUND AND VIBRATION IN BUILDINGS 



281 



will be defined as loss factor. For most materials this number is 
10-30 per cent, and does not depend upon the loading. It is also 
important that all electrical constants be almost independent of the 
amplitude of the alternating force. Materials such as cork or rubber 
are used for insulating machines, especially those of variable speed, 
and for insulating building structures. The values obtained with the 
apparatus described permit comparison of the materials and selection 
of the most suitable ones for specific purposes. 

ELECTRICAL MEASUREMENT OF VIBRATIONS 

Vibrations of buildings belong to the category of structure-borne 
sounds, except that the frequencies are lower. They lie in the neigh- 

vertical component 







FIG. 13. Vibrations in different floors of a building (solid curves: 
velocity amplitude, broken curves: frequency range). 

borhood of the lower auditory limit. Therefore, the instrument for 
measuring them must differ somewhat in construction from sound 
pick-ups, although many electro acoustical analogies may be made 
use of without difficulty. In exemplification, the essential features of 
the new vibration meter constructed at the Institute will be described. 
As a rule purely mechanical seismometers are used, which measure 
either the displacement or the acceleration of the ground as functions 
of their resonance frequencies. One type employs low frequencies 
and the other high frequencies. Similar principles would apply to 
acoustics, where, however, we measure the velocity amplitude or the 
pressure. To determine the velocity amplitude, the vibrations of 
the mechanical measuring device must be controlled by a resistance, 
not by a mass or compliance, as is usual. The solution seems to 



282 



E. MEYER 



[J. S. M. P. E. 



be impossible by purely mechanical means, but it can be done by 
electromechanical apparatus. Fig. 10 shows the principle, Fig. 11 the 
apparatus itself. A pot-shaped magnet serves as a heavy mass, and 
is supported by a lever. By means of a spring a low resonance fre- 
quency of about one cycle per second is obtained. A coil fixed to a 
base plate, and which vibrates with the ground, dips into the circular 
air-gap of the magnet. For oscillations of the ground somewhat 
higher than the resonance frequency of the system, the heavy magnet 

remains at rest. Consequently, 
the relative movement between 
the magnet and the ground is at 
the same time a measure of the 
absolute velocity of the ground. 
The emf . in the coil thus deter- 
mines the velocity of the ground, 
at least within the frequency 
range 1.5-100 cps. The response 
curve (Fig. 12; with amplifier) 
is obtained by means of an oscil- 
lation desk consisting of a plate 
and a spring of variable length, 
the desk being made to oscillate 
by an electromechanical feedback 
device. The amplitude of the 
desk with which the calibration 
is performed is read with a 
microscope. 

There are three measuring sys- 
tems, one for the vertical move- 
ment, two for the horizontal motions. The emf. of the system is am- 
plified and its average value is measured, with a time-constant of 0.2 
second. In this manner the intensity of the vibrations is determined. 
For measuring the frequency, the output of the amplifier drives a 
polarized relay, which establishes alternately a condenser charging 
and discharging circuit. The discharge current indicates the fre- 
quency of the vibrations, or rather, in the presence of several fre- 
quencies, the frequency range. In such manner the amplitude and 
frequency (average values) of vibrations of a building, caused by the 
traffic in the streets, for instance, are determined. With the appa- 
ratus described about ten thousand single measurements have been 




height of f loon relative to the top floor 



FIG. 14. General distribution of 
vibrations in buildings (1-9) of differ- 
ent height ; horizontal component. 



Mar., 1937] SOUND AND VIBRATION IN BUILDINGS 283 

made in Berlin in various buildings. Fig. 13 is an example of these 
measurements. The abscissas represent the time (about x /4 hour), 
the ordinates the velocity amplitude (full-line) or the frequency 
range (dashed-line). Both the vertical and the horizontal components 
increase with the height of the floor. It is very remarkable, and has 
been observed in many buildings, that the frequency range in the 
vertical component is filtered out. In the uppermost floor only a 
small range of frequencies is found: namely, the resonance periods 
of the ceilings, whatever the spectrum of the traffic vibrations may 
be. The building is a mechanical wave-filter in the vertical direc- 
tion; but this effect has never been observed in the horizontal. 
Moreover, for the horizontal velocity component the form of the 
vibrations was very similar in all buildings. Fig. 14 represents a 
series of measurements in various buildings of different heights; the 
abscissa is the height relative to the uppermost floor, and the ordinate 
gives the relative amplitudes, also referred to the value at the highest 
floor. It is seen that horizontal vibrations of two types are present, 
the fundamental (building as a whole) and the second harmonic (one 
node). It is very surprising that this knowledge can be obtained by 
a statistical excitation such as represented by traffic. 

In the foregoing only a few examples have been selected to show the 
relations between mechanics and acoustics in building vibrations; but 
they show clearly that not only all the measuring devices, but 
also the methods of treating the problems, have been helped much by 
electrical engineering. It seems quite necessary to make electrical 
analogies. Only by very exact physical investigation of the individual 
factors can it be hoped to accomplish a survey of the complex propa- 
gation of sound and vibration in buildings. 



THE USE OF VISUAL EQUIPMENT IN ELEMENTARY AND 
SECONDARY SCHOOLS* 



C. M. KOON** 



Summary. Encouraged by a grant from the American Council on Education, 
the United States Office of Education launched the National Visual Instruction 
Survey in January, 1936, to determine (1) the nature of the visual and auditory 
equipment owned by elementary and secondary schools in the United States, (2) the 
extent of its use, and (5) ways in which national agencies can facilitate the use oj 
visual and auditory aids for instructional purposes. The author of this article 
directed the survey, and Mr. Allen W. Noble assisted him. Reports were received 
from approximately 9000 school systems covering 95 per cent of all cities with a 
population of 5000 or more, and a fair percentage of the rural school districts. The 
paper summarizes the results of the survey. 

Dr. John W. Studebaker, U. S. Commissioner of Education, is of 
the very definite opinion that schools need audio-visual aids if they 
are to be efficient and up to date. Yet, the recently complicated 
National Visual Instruction Survey conducted by the Office of Edu- 
cation disclosed the fact that, of the 82,000 schools covered by the 
survey, representing a total enrollment of some 17,000,000 pupils, 
only 50,012 units of radio and visual projection equipment were re- 
ported as being owned by these schools. The schools and school 
systems covered by the survey were mainly located in urban centers. 

The survey was conducted for the purpose of determining, first, 
the frequency of use of the various audio-visual aids in each school or 
school system reporting; second, to secure a list of all audio-visual 
equipment owned by each school or school system; and, third, to 
arrive at an analysis of the problems met in the administration of a 
program of audio-visual education. 

It is estimated that the 33,000,000 persons enrolled in America's 
public and private schools spend about 198,000,000 hours per day 
in school, as compared with 195,000,000 hours per week spent in our 

* Received January 19, 1937; presented at the Fall, 1936, Meeting at Roches- 
ter, N. Y. 

** Office of Education, U. S. Dept. of the Interior, Washington, D. C. 
284 



VISUAL EQUIPMENT IN SCHOOLS 



285 



theaters. Yet a breakdown of the equipment owned as disclosed by 
the national survey reveals that throughout the schools covered there 
are but 9304 motion picture projectors, of which only 793 are equipped 
for sound; 11,500 radio receiving sets; and 941 central sound sys- 
tems. There are some 280,000 elementary and secondary schools in 
the United States more than eighteen times the number of film 
theaters and more than 438 times the number of broadcasting stations. 
It would appear from this that while mechanical improvements in 
the commercial entertainment field march steadily forward, the use 



10 20 JO 



in Perceitrj| 

50 60 TO 10 




ENROLLMENT 10,000 UP 
LEGEND 



FIG. 1. 



OFTEN SOME NEVM 

Extent to which motion pictures are used in elementary and second- 
ary schools. 



of these improvements, so necessary for better transmission of modern 
educational ideas in a modern way, lags far behind in our Nation's 
schools. Obviously, audio- visual education can not expect to make 
much headway with less than one unit of equipment per school. 
Mechanical and non-mechanical aids to teaching technic for the 
intellectual development of the young people of this country is of 
vital importance. Present-day social problems arising out of the 
complexities of our times will be effectively solved only if the teacher 
is given every aid created by the mind of man to combat them. 

The survey brought to light one very significant point. Almost half 
of the schools reporting were not electrically serviced to accommodate 



286 C. M. KOON [J. S. M. P. E. 

mechanical visual equipment. In spite of this fact, two-thirds of the 
school systems reported that they make some use of visual equipment 
in their teaching. 

A glance through the total inventory of mechanical visual aids 
shown as being owned by the schools circularized is revealing, to say 
the least. The equipment in Table I was listed as owned by the 
schools reporting : 

TABLE I 

Total Equipment Owned by Schools 

17,040 Lantern-slide projectors 

3,007 Still-film attachments 

2,733 Slide-film projectors 

2.073 Micro-slide projectors 
2,720 Opaque projectors 

6.074 16-mm. silent motion picture projectors' 
458 16-mm. sound motion picture projectors 

3,230 35-mm. silent motion picture projectors 

335 35-mm. sound motion picture projectors 

11,501 Radio receiving sets 

841 Centralized radio-sound systems. 

Using owned equipment as a basis for calculations, the startling 
inadequacy of audio-visual equipment in the Nation's schools is 
emphasized by dividing this list by the approximate total of more than 
81,000 schools from which it was derived: 

TABLE II 

Equipment Owned per School 

0.210 Lantern-slide projectors 

. 037 Still-film attachments 

0.034 Slide-film projectors 

0.026 Micro-slide projectors 

. 034 Opaque projectors 

0.075 16-mm. silent motion picture projectors 

. 00565 16-mm. sound motion picture projectors 

. 040 35-mm. silent motion picture projectors 

. 004 35-mm. sound motion picture projectors 

. 142 Radio receiving sets 

. 0104 Centralized radio-sound systems 

Mathematical limitations forbid further calculations to determine 
the fraction of unit of equipment per 17,000,000 children served by 
these schools. 



Mar., 1937] VISUAL EQUIPMENT IN SCHOOLS 287 

These figures are, of course, extreme. In addition to the owned 
equipment many pieces of apparatus were reported as rented or bor- 
rowed. Many radio sets, for instance, installed in schools are the 
personal property of teachers and students; many of them belong 
to student clubs. Other equipment actually in the school may be the 
property of persons who leave them there for demonstration purposes. 

In addition to mechanical devices, non-mechanical aids such as 
objects, models, and specimens; wall maps; charts and graphs; 



f prjiwd if. ff.C.Mj(ft 

K> 50 



f H.MSUIPS AND STIUFIIMS 



FIG. 2. Extent of use of all audio-visual aids in junior high schools of 750- 
2499 enrollment group. 

mounted pictures; posters; cartoons; and stereographs were shown 
as being in wide use by the survey. All kinds of aids in this category 
showed strong resemblance in their relative frequency of use. First, 
being inexpensive and for the most part easy to obtain their use is 
more widespread and more frequent than mechanical aids. Second, 
the extent to which they are used progresses in direct proportion to 
the size of the school system using them. In this group, however, the 
extent of use of such aids ranged from the almost universal use of wall 
maps to the comparatively limited use of stereographs. 

Mechanical audio-visual aids, which include motion pictures, 
lantern-slides, slide-nuns, radio programs, and phonograph records 
are, quite understandably, not employed as widely in education as 
are the simpler non-mechanical aids. However, the survey brought 



288 C. M. KOON [J. S. M. P. E. 

to light a rapid rate of growth in the school employment of both 
radio programs and motion pictures. In the larger systems, a greater 
and more uniform use of all audio-visual aids was reported. 

Lumping together mechanical and non-mechanical aids, it is 
interesting to analyze the frequency of use, as reported by the survey, 
of all the various aids in the different pupil groupings. For instance, 
the use of wall maps shows first place in use from the smaller schools 
through the largest. The use of charts and graphs comes second in 
the smaller institutions until the 2500 or more level is reached, be- 
yond which posters and cartoons rank second, charts and graphs 
third. Fourth place went to mounted pictures throughout the 
various levels. Likewise, objects, specimens, and models were uni- 
formly fifth in rank. Next in order of use came phonograph records. 

The use of radio programs shows wide variation among the various 
enrollment levels. In the smallest schools, that is, those serving from 
one to 750 students, the use of radio ranks seventh. In the next 
level (750 through 2499) radio instruction falls to ninth place. In 
schools serving from 2500 to 10,000, radio is next to last in frequency 
of use, and in the largest schools it ranks last. 

While the use of radio, ranking seventh in the smallest schools, 
is but fifteen per cent, its frequency of use in the largest schools, 
where it ranks last, is twenty-two per cent. Thus it will be seen that 
while radio appears to drop as the pupil level increases, actually it is 
more widely used in larger schools although other audio-visual aids 
outrank it in frequency of use. 

The use of motion pictures ranks eighth in all enrollment levels. 
Next in the smallest schools is the use of lantern-slides, although this 
device ranks seventh from the 750-student level up. Similarly, 
stereographs show next to last in the two lower levels, while their 
use ranks ninth in the upper brackets. Least of the visual aids to 
be used are slide-films and still films, with the exception of the largest 
schools, where this method outranks radio programs by a fraction of 
a per cent. 

Table III shows the frequency of use of the various audio-visual 
aids among the different enrollment levels. 

A still further analysis of the data shows an increasing rate of 
use of the various audio-visual aids as the grade levels advance 
until the senior high school is reached. In the main, senior high 
schools do not make as much use of audio-visual aids as do the junior 
high schools. Table IV, for motion pictures, illustrates this point. 



Mar., 1937] 



VISUAL EQUIPMENT IN vScnooLS 



289 



While the foregoing percentages of frequency are rather gratifying 
in view of the low percentage of unit of equipment per school, the 
survey forces one to the unavoidable conclusion that our schools, 



TABLE III 



Analysis of Frequency of Use of Various Audio- Visual Aids in Education 

(Per Cent) 



Pupils Enrolled > 


Under 7 50 


750-2499 


2500-9999 


Over 10,000 


Reported as Using > 


Often Some 


Often Some 


Often Some 


Often Some 


Wall maps 


69 26 


76 20 


76 19 


80 16 


Charts and graphs 


52 42 


56 38 


56 39 


64 32 


Posters and cartoons 


50 45 


45 52 


58 39 


61 38 


Mounted pictures 


43 49 


47 48 


50 45 


61 38 


Objects, specimens, etc. 


37 55 


43 52 


46 48 


59 40 


Phonograph records 


34 45 


42 48 


44 47 


54 43 


Lantern slides 


11 36 


26 50 


29 54 


48 41 


Motion pictures 


15 36 


26 46 


28 49 


41 47 


Radio programs 


15 47 


18 51 


18 57 


23 61 


Stereographs 


10 24 


17 32 


19 43 


25 47 


Slide-films and stillnlms 


5 20 


13 29 


13 37 


23 44 



both public and private, are poorly equipped to get the ultimate 
results attainable through a more widespread use of mechanical and 
non-mechanical means of graphically and entertainingly impressing 
upon the mind of the student vital facts in the study of science, 
geography, history, social science, health, English, nature, commerce, 



TABLE IV 



Use of Motion Pictures on the Different Grade Levels 
(Per Cent) 



Pupils Enrolled > 


Under 750 


750-2499 


2500-9999 


Over 10,000 


Reported as Using > 


Often Some 


Often Some 


Often Some 


Often Some 


Primary (Grades 1, 2, 3) 
Intermediate (Gr. 4, 5, 6) 
Junior high (Gr. 7, 8, 9) 
Senior high (Gr. 10, 11, 12) 
Average use all grades 


11 31 
14 34 
17 38 
17 41 
15 36 


15 45 

23 45 
33 43 
31 50 
26 46 


17 49 

25 49 

36 46 
32 52 
28 49 


26 54 
40 48 
51 39 
44 46 
41 47 



and industry; indeed, practically all subjects included in the school 
curriculum. 

The question why are not these important audio-visual aids to 
teaching technic in more widespread use in the schools and school 
systems throughout the United States? logically frames itself in 



290 C. M. KOON [J. S. M. P. E. 

the mind of anyone studying the findings of the survey. It is interest- 
ing, then, and important to explore the reasons given by those in 
charge in the various school units as to why audio-visual equipment 
has been so slow in getting a foothold in our schools. 

The greatest handicap reported was lack of sufficient budgetary 
provision for the work. Next was the fact that the schools were un- 
able to get the proper aids in the classroom when needed most. Third 
greatest difficulty was the declaration that the teachers were in- 
sufficiently trained in the use of visual aids. The fourth complaint 
was that the available aids fail" to cover the course of study adequately. 
Fifth was a lack of understanding of the value of visual aids. The 



ATTENDANCE 

750-ZVW ZSCO-WT 




R.IPOR.TJ 
V20I 1113 591 ISt 

FIG. 3. Schools encouraging attendance at selected motion pictures. 



sixth difficulty recorded was the lack of information on sources of 
desirable films and other aids. 

A space was left open in the questionnaire so that city and county 
public school superintendents and principals of private high schools 
might express any other difficulties not specifically mentioned in the 
questionnaire. In the survey's compilation of reasons given here it 
is interesting to note that many superintendents and directors wrote 
that they were personally in favor of the use of visual aids, and that 
their faculty was reasonably trained in the use of such equipment, 
but that they were unable to convince the school board and others 
in charge of fiscal policies of the merits to be found in the use of visual 
aids. This can be termed lack of understanding of the value of such 
aids, and materially affected the budgetary provisions. 

In an effort to determine what means might be set up to offset 
these difficulties, the survey listed several suggestions in the question- 
naire. By far the greatest interest was centered upon the formulation 



Mar., 1937] 



VISUAL EQUIPMENT IN SCHOOLS 



291 



of some plan whereby equipment could be purchased with the assis- 
tance of some Federal agency. By this was not meant merely to 
purchase the equipment at reduced costs or on time payments, but 
that there was a need of assurance that the equipment so purchased 
would be standard as to quality and size, in order that films could be 
obtained for it. With the development of a Federal purchase plan, 
such as the Electrical Home and Farm Authority is now formulating, 
it will be possible to have the equipment inspected and certified by the 



15 



























1 1 1 








Mlin i 


SCIENCt 


Ttm 


HI J TO*. 


MCMi' NIACTMI IM4UMJ HATURl IcOMV.lRCtl JMtir HHMClJ0HCi.TIj ormtu 



FIG. 4. Relative volume of usage of educational films by subjects (17,228 

listings). 



Bureau of Standards, which will do much, the survey believes, to 
offset some of the prejudices recorded. 

Many instances were disclosed where obsolete and odd-sized equip- 
ment had been sold to schools and school boards. Very little use can 
be made of this equipment. 

Demonstration lessons in the schools by visual instruction experts 
received the second heaviest vote. Next came the expression of a 
need for lesson plans to aid in the correlation of visual aids with the 
course of study. Ranking fourth was the need for additional motion 
pictures produced for instructional purposes. Next in order of favor 
was the establishment of visual instruction centers where courses 
could be offered. On this point there seemed to be a difference of 



292 C. M. KOON [J. S. M. P. E. 

opinion as to the most expedient manner of bridging this gap, as 
many of the teachers in the schools have had little or no training 
in the use of visual aids. 

Three courses are open in this connection. First, that the teacher 
be required to attend a university or college and obtain such training. 
Several states already require this of new teachers, but the main 
problem lies in training those who have completed the requirements 
imposed prior to this time and who are now teaching. The second 
alternative is to require the teacher to take a course in the use of 
visual aids from an extension division of one of the many universities 
offering such courses. The third means of training the present staff 
is to employ supervising teachers of visual instruction whose duty will 
be to work with the teachers and show them how to select the films, 
slides, etc., to be used by the entire school. Several school systems 
have reported that the latter course seemed the most expedient and 
flexible, and reported the successful use of this means. 

The next development desired by those in the field, the survey 
revealed, was the establishment of some group or groups to give expert 
evaluation of educational films and other visual aids. Many superin- 
tendents and directors of visual education reported that they were too 
frequently disappointed in the contents of the films, and that the 
descriptions supplied were inadequate and often misleading. It is 
interesting to note that almost without exception the entire platform 
of the proposed American Film Institute was approved by the 
respondents in the National Visual Survey. 

Following the logical why as to the scarcity of audio-visual equip- 
ment throughout the schools at the present time, comes the even 
more logical what is to be done both by educators and producers of 
equipment to overcome this deficiency? 

In general, it may be said that the most important need in the 
development and extension of the use of visual aids in education is 
more cooperation and better understanding between motion picture 
and other visual-aid producers and manufacturers, on the one hand, 
and educators and socially minded groups, on the other. The former 
need to become more fully aware that this vast potential market 
must be developed by long-time planning on a sound economic 
basis. The latter need to appreciate more fully the importance of 
the technical, professional, and business ability of people in the 
industry, and the fact that both must work together on a public 
spirited but sound economic basis. More specifically, the following 



Mar., 1937] VISUAL EQUIPMENT IN SCHOOLS 293 

suggestions are made for the further development of the field of 
visual education: 

(1) Further improvements in the 16-mm. projector to make it even easier to 
operate and essentially "fool-proof." 

(2) More uniformly good reductions from 35- to 16-mm. prints. 

(5) The Society of Motion Picture Engineers vigorously discourage attempts 
to sell schools freak and impracticable types of visual aids, as well as splendid types 
of aids that get out of order easily. 

(4) The Society of Motion Picture Engineers to work out minimum per- 
formance standards for 16-mm. projectors intended to show either silent or 
sound films in semi-dark rooms. 

(5) The further development of deferred payment plans for major units of 
visual equipment. Perhaps the Federal Government should participate in some- 
what the same way as it has in encouraging farmers to purchase electrical farm 
and home equipment in the Tennessee Valley Authority and Rural Electrification 
projects. 

(6) A more efficient distribution system for educational films and equipment. 

(7) Closer cooperation between industry and educators in determining the 
content and treatment of subjects in new films and other forms of visualization. 

(8) More teacher-training courses in the technic of teaching with motion 
pictures and other visual aids, as well as more courses in photoplay appreciation. 

(9) A vigorous publicity campaign of a high order, reaching all the 9000 key 
people listed in the new National Visual Education Directory* just issued by the 
American Council on Education, Washington, D. C. 



DISCUSSION 

MR. MITCHELL : Some of you recognize that the trend recently has been toward 
definite recognition of motion picture applications in teaching. A number of 
teachers have been reluctant to use motion pictures, particularly sound pictures, 
because they rather felt that it eliminated the personal touch, or perhaps tended 
to eliminate the teacher entirely. 

Enough experience has been accumulated under carefully controlled conditions 
to prove beyond question to even the most skeptical that the use of sound motion 
pictures shows educational results far in advance of any other medium. They 
have been tested under such conditions that no argument remains. They have 
been recognized officially by Government departments, by the President, and 
by a Council sponsored and patroned by Mrs. Roosevelt, the President's mother. 

Such official recognition naturally has an effect upon school officials who are 
politically minded. It is fortunate that we have reached this stage, and those 
close to the field have recognized how very rapidly it is progressing. 

* Copies of the National Visual Education Directory listing the addresses of 
9000 visual leaders in local school systems and the equipment owned, may be 
purchased from The American Council on Education, Washington, D. C. 



ORGANIZATION AND WORK OF THE FILM LIBRARY OF 
THE MUSEUM OF MODERN ART* 



JOHN E. ABBOTT** 

Summary. Until last year no organization existed for preserving films of out- 
standing merit or for arranging for their distribution and study by those interested 
in film as living art and in its history and development. A grant from the Rocke- 
feller Foundation and private gifts permitted the Museum of Modern Art to establish 
such a Film Library in June, 1935, under the presidency of John Hay Whitney, 
with Will H. Hays, Chairman of the Advisory Committee. 

The functions of the Film Library are to trace, obtain, and preserve important 
films, American and foreign; to edit and assemble such films into programs for 
educational and non-commercial exhibition; to arrange notes and critical appraisals 
of them; to assemble a library of books and data on the films; and otherwise to make 
available information concerning their artistic, dramatic, and historical aspects to 
all who may be seriously interested. The series for 1936 consists of (1) The Develop- 
ment of Narrative (1895-1911); (2) The Rise of the American Film (1912-17}; 
(3) D. W. Griffith (Intolerance); (4) The German Influence; (5) The Talkies. 

Although the film has been with us for nearly half a century now, 
until last year no organization existed anywhere for preserving 
motion pictures of particular merit or interest. People could read 
and talk about pictures of the past, such as Sarah Bernhardt's Queen 
Elizabeth, which Mr. Zukor brought over from France and by doing 
so established the feature film; but people could not see Queen 
Elizabeth; nor, for that matter, could they have a look at The Jazz 
Singer, which had an even more overwhelming influence upon the 
course of the film's history. It was necessary for us to depend upon 
our memories, our very inadequate and inaccurate memories of the 
film, if we wanted to form an estimate of the changes, the advances 
that have occurred since Al Jolson's mammy song rang the knell of 
the silent pictures. 

That it would be highly desirable to preserve outstanding films 
had been generally agreed by everyone but that it was practicable 
seemed highly doubtful. It hardly lay within the scope of the film 

* Presented at the Spring, 1936, Meeting at Chicago, 111. 
** Directort, Museum of Modern Art Film Library, New York, N. Y. 
294 



ORGANIZATION OF FILM LIBRARY 295 

industry itself to tackle the job: for one thing, it would be an ex- 
pensive task, and one without appropriate financial returns. What 
is more to the point, however, is that no man, or body of men, can 
look both forward and backward at the same time. The energies 
and talents of the industry are necessarily canalized into today's and 
tomorrow's activities; a quite different set of energies and talents 
are needed for the backward perspective, for the collecting or li- 
brarian's habit. Also, as it is notorious that artists are often the 
worst judges of their own work, it was to be suspected that the 
creative ability of the film industry as a whole would be inadequate 
for selecting from its vast output what among that output had been 
the most memorable and best worth preserving. 

The first attempt to form a library of excellent films seems to have 
been made in Denver, Colorado, in 1922, by George William Eggers, 
the Director of the Denver Art Museum. Later attempts to revive 
meritorious films followed in many large cities all over the world, 
but seldom endured for long. After ah 1 , prints wear out, and receipts 
from occasional revivals hardly justify making new prints. Only 
one or two especially famous films continued to pop up from time 
to time. All the rest of the huge output of the industry vanished 
once its relatively brief period of box-office productivity was ended. 
The probabilities were, therefore, that ah 1 the films, bad and good, 
made since the beginnings of the industry, must continue to lie for- 
ever unseen in their various vaults and warehouses. In the course 
of time they would deteriorate and eventually perish; therefore, to 
future generations the nature of films of the past would continue to 
be a matter of legend, conjecture, and mystery. In other words, the 
film was condemned to enjoy only a transitory life even the best films 
were condemned to enjoy only a momentary space in the sun and there- 
after to become revered memories like the very famous plays of the 
commedia delVarte about which, actually, so little is known although 
so much is said. It seemed a great pity. 

Now, if anything were to be done to create a museum of the film, 
it did seem obvious that a singularly appropriate institution to under- 
take the work was the Museum of Modern Art, in New York. 
Founded in 1929 it has energetically concerned itself with contempo- 
rary art in all its aspects, from architecture to photography, from 
painting to typography. And since the foundation of the Museum 
of Modern Art, the director and the trustees had always planned a 
department of films. Yet, before the Museum of Modern Art could 



296 J- E - ABBOTT [J. S. M. P. E. 

approach the task, three things were needed. One was to ascertain 
whether there existed a serious interest in the film as living art. 
Another need was money, to create and maintain a film library; 
and the third was the cooperation of the film industry. 

Inquiry proved that colleges and museums all over the country 
were anxious for material to make possible a serious study of the film. 
Thereupon, a scheme for the creation and operation of a film library 
was drawn up such as would enable the motion picture to be studied 
just as, for example, mediaeval sculpture or contemporary drama 
is studied. A grant from the Rockefeller Foundation and certain 
gifts of money from private individuals provided the necessary funds 
to start work. The Museum of Modern Art Film Library came into 
existence in June, 1935. Its officers are John Hay Whitney, President; 
Edward M. M. Warburg, Treasurer; and John E. Abbott, Vice- 
President. The last named was appointed Director, and Iris Barry, 
Curator. Later, an advisory committee was formed with the follow- 
ing members : 

WILL H. HAYS, Chairman 

JULES BRULATOUR 

STANTON GRIFFIS, a trustee of Cornell University 

DR. IRWIN PANOFSKY, Professor of Fine Arts at the Institute for Advanced 
Studies at Princeton, N. J. 

Dr. DAVID H. STEVENS, Director for the Humanities of the Rockefeller Founda- 
tion 

IRVING THALBERG 

The Film Library then set to work. There was no precedent for 
what it hoped to do. Its plans were : 

(1) To compile and annotate a card index of all films of interest or merit of 
all kinds produced since 1895, both American and foreign. 

(2} To trace, secure, and preserve the important films of each period since 
1893. 

(5) To edit and assemble these films into programs for exhibitions in 
New York and throughout the country by colleges, museums, and local 
cultural organizations. 

(4) To compose program notes on each exhibition, which would include a 
critical appraisal of the films and would aid the student in appreciation of the 
medium. 

(5) To assemble a library of books and periodicals on the film, and other 
historical and critical material including the vast amount of unrecorded data 
still in the minds of men who developed the film. If the history of the formative 
period is to be preserved, it was necessary to obtain this information at once; 
otherwise, it would be irrecoverably lost at the death of these men. 



Mar., 1937] ORGANIZATION OF FILM LIBRARY 297 

(6) To assemble and catalog a collection of film "stills." 

(7) To preserve and circulate the musical scores originally issued with the 
silent films, and to arrange musical scores when the original has been lost (sheet 
music or phonograph records) to be circulated with the silent programs when 
needed. 

(8) To act as a clearing house for information on all aspects of the film, and 
to maintain contacts with all interested groups, both in America and abroad. 

(9) To make available sources of technical information to amateur makers 
of film. 

(10) To publish a bulletin containing articles and illustrations for furthering 
the appreciation and study of the motion picture and to make known the Film 
Library's activities. 

We immediately approached the film industry in this country, 
first, through Motion Picture Producers & Distributors of America, 
Inc., and then, individually, the executive heads of producing com- 
panies. What we asked of them was, of course, the use of certain 
films they owned. The Film Library was in a position to pay the cost 
of restoring negatives, prints, and circulation. There were in- 
numerable difficulties to be ironed out, such, for instance, as the 
complications arising in the case of a film that has been remade by a 
firm other than the one that originally produced it. And, of course, 
the conditions under which the Film Library would circulate films 
had to be defined and agreed upon. Despite the obvious legal and 
technical difficulties, once our objects were made clear the fullest 
cooperation was forthcoming, and in almost every case the films 
we asked for have been made available to us. Also, early in the 
Library's existence we were fortunate in acquiring the collection of 
early films and other material that the late Jean A. LeRoy had 
amassed. 

Our first series of films for circulation was released in January, 
1936, under the title of a "Short Survey of the Film in America from 
1895 to 1932." It was planned as a first-year course, or survey, 
which would provide a groundwork for a more voluminous series in 
the following year. It was composed almost exclusively of American 
films, not only because of the predominant part this country holds 
in the film field, but also because historically the native film has been 
on the whole much less seriously considered and appreciated than it 
deserves. The idea seemed to have become prevalent that foreign 
films were art, but hardly the domestic film. This was a point of 
view that seemed untenable. This preliminary series of films con- 
sisted of five programs, each, roughly of two hours' duration, issued 



298 J- E. ABBOTT [J. S. M. P. E. 

one a month. The first program was entitled The Development of 
Narrative, and covered the years 1895 to 1911, beginning with the 
Edison Company's brief but business-like Execution of Mary Queen 
of Scots and ending with Sarah Bernhardt in Queen Elizabeth. It 
also included the two deservedly famous items, A Trip to the Moon 
and The Great Train Robbery. 

The second program, called The Rise of the American Film, included 
early films by D. W. Griffith and Mack Sennett, and an early Western 
from the Ince studio, with Wm. S. Hart; it concluded with the 
celebrated "vamp" film A Fool There Was. The third program was 
devoted entirely to D. W. Griffith's Intolerance. The fourth program 
was called The German Influence and included the late F. W. Murnau's 
first American-made picture, Sunrise. The last program traced the 
development of the talking film, and included two scenes from The 
Jazz Singer one of them, of course, being the one in which Jolson 
sings his mammy song and the whole of All Quiet on the Western 
Front. Last item on this program, and in the series, is Steamboat 
Willie, the first Mickey Mouse release. 

Where, one may ask, were these films shown? It would be tedious 
to give a list of the exhibiting institutions ; but so far they have been 
fifty in number, and include eleven colleges, twelve universities, 
fifteen museums, two art schools, and ten adult cultural organizations, 
located in twenty states. 

Each program is prefaced by a rolling title of fact and comment, 
and other titles giving the necessary historical data preface each single 
film. Scores for piano have been arranged as accompaniments for 
all the silent programs, and the music is sent out in advance to give 
time for rehearsal. But what we feel is the most important feature of 
the programs, except for the films themselves, is that carefully writ- 
ten printed program notes are also provided for each member of all 
the audiences seeing the films we circulate. These program notes 
are intended, on the one hand, to convey information about the 
history of the motion picture, and, on the other hand, to help students 
both to gain a thorough grasp of its development and to arrive at a 
more analytical and more detached attitude to the film generally than 
is possible when films are shown currently for entertainment only. 

I think there is little question that anyone who has seen the five 
programs I have briefly described has now a good grounding in the 
history of the art and has acquired a totally new respect for, and a 
much profounder understanding of, the medium as a whole. Those 



Mar., 1937] ORGANIZATION OF FlLM LIBRARY 299 

of us at the Film Library have to admit that we have gained a great 
deal ourselves. Considerable misinformation about films has ac- 
cumulated over a period of time, and this was inevitable so long as 
what we knew about them was largely hearsay and recollection. It 
is quite an experience to review the history of the film again. For 
instance, how often has it been said that D. W. Griffith invented the 
close-up? Actually one very early film, the famous John C. Rice- 
May Irwin Kiss, turns out to have been taken entirely in close-up 
in 1896. That knowledge makes it even more interesting to see, when 
we come to D. W. Griffith's early films, that while he did not invent 
the close-up, he most definitely discovered how to use it. And then, 
again, it is commonly believed that the moving camera, tracking 
shots, came in with the German films after the War. But in our 
third program in Griffith's Intolerance there are startlingly effective 
and long-sustained tracking shots, where the camera swings across 
the great outer courtyard of a Babylonian palace and is then moved 
slowly forward. 

There is a real thrill, comparable, I suppose, to the thrill with which 
an Egyptologist uncovers a new burial chamber on the Nile, in being 
able now, for the first time, actually to trace the innumerable in- 
fluences, packed into an unbelievably short span of years, that have 
produced the films we see today. It gives a real sense of satisfaction 
to be able to trace a clear line of descent from the old Max Linder 
comedies, through the Keystone comedies to the work of Ren Clair; 
or gradually to see the connection between the technic that Porter 
hit upon in The Great Train Robbery being carried a stage further in 
the extraordinarily free cutting of Intolerance; and then, scientifically 
analyzed and applied, to see the same technic applied to the much 
later Russian films. We believe that being able to review and study 
the history and development of film technic, to view the medium in 
perspective, will prove to be a tremendous benefit to future technicians 
of the film. 

Our work has only begun ; many of our difficulties still lie before us. 
We have been much encouraged by the response of the academic 
world to our activities, and by the generous cooperation of the old- 
timers and the newcomers in the industry itself. We most urgently 
need the help, the criticism, and the encouragement of all who belong 
to the industry in our effort to preserve the outstanding films of the 
past, to make their development understood, and to create a con- 
sciousness of tradition and history within the art of the film. 



300 J- E. ABBOTT [J. S. M. P. E. 

(A reel of three subjects chosen from the first series of films assembled by the Film 
Library lias projected at the close of the paper.} 

The following represents typical material supplied with the films 
when they are distributed : 

A SHORT SURVEY OF THE FILM IN AMERICA CIRCULATED BY 
THE MUSEUM OF MODERN ART FILM LIBRARY 

PROGRAM I 

The Development of Narrative: 1895-1911 

The Execution of Mary Queen of Scots 1895 

(Produced by The Edison Co. Directed by William Heiss. Cast: Unknown.) 

Once people had overcome their amazement that the moving picture really 
moved, the new invention was appreciated because it could record the present 
topical views of Fifth Avenue or a French railway station were much enjoyed 
and because it could recreate the past. This ruthless little Edison film was made 
for the peep-show, or kinetoscope. After 1896, films were commonly shown on 
screens. 

Wash Day Troubles 1896 

(Directed by Edmund Kuhn. Cast: Mrs. Edmund Kuhn and others.) 

The film could do more than record and recreate : it could invent new stories. 
Comic incidents like this were improvised from the earliest days, and developed 
later into slapstick comedy. 

// was after the period now under review that the American film developed its own 
characteristic expression. During the early years foreign films contributed much to 
the repertory of the cinema in America. Trick films and historical films from France, 
spectacles from Italy and melodramas from England were widely shown. 

A Trip to the Moon 1902 

(Produced and directed by George Melies. Story, design, scenery, effects by 
George Melies. Cast: George Melies, dancers from the Theatre du Chatelet, 
acrobats from the Folies Bergeres.) 

Melies made hundreds of films between 1896 and 1914. His experience as an 
illusionist and magician helped him to exploit the camera's resources, so that he 
quickly hit upon stop-motion photography for magical transformations and 
disappearances, used a close-up as early as 1896, and made the first film by 
artificial light. His influence can be traced in innumerable films but most 
markedly in French advance-guard work like Cocteau's The Blood of a Poet. 

Clearly, Melies regarded the film literally as a series of tableaux vivants, 
though he achieved a continuity by overprinting the end of each scene on the 



Mar., 1937] ORGANIZATION OF FlLM LIBRARY 301 

beginning of the next. As in most early narrative films, the scenery is disposed 
and the characters move horizontally as on a stage in view of an audience. 

The beautiful painted backcloths, movable scenery, and stylized props Melies 
himself designed, and executed in grisaille: he was an enthusiastic student of 
Prud'hon, Delacroix, and other Romantic painters. M61is notably developed 
screen narrative by ransacking literature for stories to tell and by telling them 
in his own energetic and imaginative manner. 

A Trip to the Moon was made when most films were only three minutes long. 
It displays the richness of Melies' invention and humor: his zooming close-up 
of the face of the moon is masterly and the whole film is charmingly unrealistic 
and gay. 

This print is incomplete: there should be thirty instead of the present twenty- 
five tableaux. Scenes of rejoicing when the explorers return to port are missing 
at the end. 

The Great Train Robbery 1903 

(Produced by The Edison Co. Directed by Edwin S. Porter. Cast: George 
Barnes, "Bronco Billy" Anderson, A. C. Abadie, Marie Murray, and others.) 

This short story in cinematography created a sensation upon its appearance 
and has become a classic of the screen. The whole feeling of the film is definitely 
cinematic, movement is employed toward and away from the camera, as well as 
horizontally in front of it. Once or twice the camera is even swung to follow 
the action. The sudden close-up with which the picture ends could, according 
to the original Edison catalog, be inserted either at the end or the beginning as 
desired. 

The story is told falteringly and the continuity is crude, yet already here is a 
feeling for the interplay and assembly of short shots which D. W. Griffith after- 
ward mastered and the Russians have since developed into modern montage. 
Porter had hit upon the principle of film editing. 

Its style is mixed. In the chase scenes there is not a trace of the theatrical: 
the dance-hall has a stage backdrop with painted stove and lamp ; in the station- 
master's office a horizontal stage-set looks onto the great outdoors. 

When issued, the film was printed on tinted celluloid yellow for the dance- 
hall, bluish green for the woods. Since tinted stock is seldom used today, it 
has not proved practicable to match the color of the original. 

Faust cl905 

(Produced by Pathe. Director and cast: Unknown.) 

This is not the Faust that M61ies made, but another version. The last hah* 
of the film is missing. Like so many of the early French pictures it was issued 
either plain or hand-colored: the coloring cost extra, and was done in bright 
stained-glass or stereopticon hues. 

It is interesting to note how, in an effort to let the audience know what the 
actors were thinking, "visions" were used as here when the tapestry on the wall 
gives place to Marguerite's memory of her meeting with Faust. 



302 J. E. ABBOTT [j. s. M. P. E. 

Queen Elizabeth 1911 

(Directed by Louis Mercanton. Acquired through the courtesy of Paramount. 
Cast: Sarah Bernhardt, Lou Tellegen and others.) 

"This is my one chance of immortality," Sarah Bernhardt said when first 
asked to act for the screen. The ability of the film to bring eminent persons of 
the day before the widest public has preserved for us here a ghostly souvenir 
of the theatrical past. It was Adolph Zukor who acquired this French picture 
and distributed it here as the first of the "Famous Plays with Famous Players" 
with which the career of the firm we now know as Paramount was begun. "Pre- 
sented by Daniel Frohman," it made its first appearance at the Lyceum Theater, 
July 12, 1912. The story of the film's introduction to and subsequent career in 
this country is well told in Will Irwin's The House That Shadows Built (Doubleday 
Doran, 1925). 

Queen Elizabeth is a photographed play rather than drama conceived in terms 
of the cinema. There is, nevertheless, a distinct shock to the spectator when 
at the conclusion, the star takes a curtain, for on the screen the player has hardly 
ever stepped out of his role. (A recent exception occurred in Escapade, with 
William Powell and Luise Rainer.) It is noticeable that the acting, which is for 
the most part decidedly statuesque, at times becomes painstakingly pantomimic, 
especially in the scenes between the Queen and the Countess of Nottingham. 
Long subtitles are also employed throughout, each of which describes carefully 
the action about to take place. 

Sarah Bernhardt's prestige surpassed that of any player now living: she could 
do no wrong, and the fact that she had consented to act for the films did much 
to diminish prejudice against the movies. The success of this film and of the 
Italian spectacle Quo Vadis helped to establish the longer feature. 

DISCUSSION 

MR. CRABTREE: Do I understand that you are obtaining the negatives of 
these films? If so, what steps are being taken to store them so as to assure 
their perpetuation? 

MR. ABBOTT: To date we have in our possession only certain very old nega- 
tives, which have been acquired from various sources. We are in touch with 
others concerning additional early films. We plan to build a vault that will be 
constructed according to the opinions of the best authorities; for the preserva- 
tion of negatives only, not positives. We hope to have arrangements ready 
within the next eighteen months. 

MR. CRABTREE: I think that if you follow what is being done in the Archives 
Building at Washington you will be getting the best information obtainable, since 
both the Film Preservation Committee of our Society, the Bureau of Standards, 
and the various film manufacturers are closely cooperating with the Archives 
personnel. 

MR. ABBOTT: If any members of the Society have bits of information or 
old documents that they feel would be of interest, we would appreciate their 
sending them to us. If they wish them back, we will photograph and return 
them. We hope to accumulate as much data as possible, and then see what 



Mar., 1937] ORGANIZATION OF FILM LIBRARY 303 

can be done toward rearranging them into some straight-line history of the de- 
velopment of the film. Any documentation would be of the utmost interest 
and assistance to us. 

MR. CRABTREE: Do you speak of films or literature? 

MR. ABBOTT: Literature as well as films, but primarily histories of certain 
reports, findings of investigations, and all sorts of printed documents that have 
been extremely difficult to locate. 

MR. TASKER: Including, I suppose, newspaper reports and records? 

MR. ABBOTT: Including everything. It may interest you to know that al- 
ready our library contains the largest collection of film books, magazines, and 
papers on the subject in existence. It is available for inspection by all who are 
interested in motion pictures. 

MR. CRABTREE: What is the fee charged for lending the films? 

MR. ABBOTT: On sending them out, we charge a fee to the college or univer- 
sity, which, at the present time, is about thirty-five per cent of the actual cost 
of preparation. The deficit is made up through an educational grant from the 
Rockefeller Foundation. 

On any one series, a five-months' program including all the program notes 
for each member of the audience, and full musical score, the charge is $125. 
That comprises, in the first series, between 50 and 60 reels, plus the program 
notes, music, and lecture material. 

MR. CRABTREE: Are single films available, or only the series? 

MR. ABBOTT : In some instances single films have been rented if the institution 
has already used the entire series as a background for study. If they want 
supplementary instruction in one subject, such as the study of the technic of 
Mack Sennett or the development of action or narration in the Westerns, we 
send additional films for their undergraduate classroom work. Naturally, we 
are not in favor of releasing single films, but would rather have them in a program 
formulated to show some specific expression. We have no intention of going 
into the distributing business for revival films. Also, the films are not available 
to groups other than established educational institutions or museums, and they 
are to be shown only to persons who are accredited members of that organization. 
They can not be booked by a commercial theater. 

MR. RICHARDSON: What has finally been determined as to the stability of 
the film ; how long will an image last? 

MR. ABBOTT: Certain experts who are interested in what we are doing have 
been cooperating with us on that problem. We had an unknown negative of 
1897, which was printed perfectly. On the other hand, we have had negatives 
of 1930 and 31 that had to be restored before they could be printed. It is very 
difficult to make a definite answer concerning the life of any negative. 

MR. RICHARDSON: If I may say this further, I have had photographs in my 
files for more than 50 years. Some of them are very badly faded while others 
are in perfect condition. Isn't it extremely likely that the same thing will be 
true of motion pictures and that the life depends upon the methods and the 
thoroughness with which they are washed? 

MR. CRABTREE: The stability depends upon many factors, including the 
nature of the film, the degree of washing, and the time, temperature, and humidity 
of storage. 



A NEON-TUBE OSCILLOSCOPE FOR THE PROJECTION 

ROOM* 



F. H. RICHARDSON** AND T. P. HOVERf 



Summary. After a brief discussion of the matter of suitable instruments and 
tools for the projectionist, for adjusting and maintaining his equipment, a neon-tube 
oscilloscope for investigating defects in the sound reproduction is described. By 
means of this instrument the presence of noises and other defects in reproduction due 
to maladjustment of the equipment can be detected, and the means of eliminating these 
distortions indicated. 

The general construction of the instrument and some of its applications are de- 
scribed. 



The advent of sound brought about many changes in motion 
picture-sound projection. Perhaps the most noticeable change is in 
the type of tools and maintenance equipment required in the modern 
projection room. True, some of the silent picture "operators" who 
boasted of twenty years' experience with one pair of plyers and a 
screwdriver as a tool kit had finally risen to the heights of an added 
hammer and perhaps even a wrench; but the projectionist of today 
who is able to take proper care of modern equipment, and make 
necessary emergency repairs, must have many tools and even some 
precision instruments. 

Unfortunately, as matters now stand, the projectionists who re- 
ceive the highest wages usually need the fewer tools, because most 
of the larger theaters paying such wages now engage maintenance 
service, and necessary repairs are made either by engineers or by 
factory representatives. 

On the other hand, in the smaller theaters, where the wages are 
lowest, the projectionists usually must do their own maintenance 
work and be able to take care of almost every kind of emergency 

* Received October 5, 1936; presented at the Fall, 1936, Meeting at Rochester, 

N. Y. 
** New York, N. Y. 

t Ohio Theater, Lima, Ohio. 
304 



A NEON-TUBE OSCILLOSCOPE 



305 



repair. These men in many cases find the outlay required for proper 
tool equipment beyond their means, and most theater managements 
fail or refuse to provide proper tools. Lack of adequate tools can not, 
therefore, be wholly blamed upon the projectionist. Having inade- 
quate means, he finds available an assortment of tools that are ques- 
tionable and very often offered at fanciful prices. This great industry 
should see to it that tools of really good quality are supplied to those 
who are required to take care of its splendid equipment, and supplied 
at reasonable prices. 

One instrument in particular, that many projectionists would like 
to have available, is a good tube-tester and combination voltmeter 
and ammeter. It has been only very recently that we have been able 
to purchase reasonably accurate equipment of this sort for less than 
a month's salary. The alterna- 
tive has in many instances been 
the construction of tools and test 
equipment by projectionists them- 
selves, or by groups organized for 
the purpose. 

About the time the cathode- 
ray oscilloscope came into prac- 
tical use many projectionists 
realized that such an instrument 
would be of great value in the 
projection room, but the cost 
acted as an effective deterrent to 
its general use. After consider- 
ing the problem, the authors came to the conclusion that it would be 
quite possible to devise an instrument that would provide at least 
a reasonably accurate record of sound phenomena and be easy to 
construct, simple to operate, and reasonable in price. 

The first experiments were with a crystal-controlled oscillating 
mirror, which projected a more or less accurate picture of the sound- 
wave, but the plan was abandoned because of the mechanical and 
optical difficulties involved. Experiments with various kinds of 
recording lamps showed excellent response to the audio signal, but 
the light emission was too feeble for successful scanning or analyzing. 
Finally a neon-tube volume indicator was chosen, after many trials 
of tubes of various kinds and characteristics. The instrument has 
been put into production at a very low price. It can be successfully 




FIG. 1. Diagram of neon-tube 
oscilloscope. 



306 



F. H. RICHARDSON AND T. P. HOVER [j. S. M. p. E. 



used by the projectionist who is willing to apply himself to the study 
of matters he must understand in any event if he hopes to give effi- 
cient results in the projection room. If the projectionist wishes to 
construct one himself, he can do so for about two and a half dollars, 
with a little ingenuity. The instrument is stable in operation, and 
with reasonable care will last indefinitely. Fig. 1 is a diagram of the 
arrangement. 

A tube of the neon type has recently been placed upon the market 
by a radio parts manufacturer that is adaptable to this instrument. 
There also is a mirror which may be attached to a motor shaft and 

used for scanning the neon dis- 
charge. The other important 
parts of the instrument are a 
variable-speed motor, with con- 
trol, and a transformer that 
will step up the amplifier out- 
put to approximately 250 volts. 
An independent amplifier may be 
used, or the regular amplifier of 
the theater sound system will 
prove satisfactory. 

Although the tube may be 
operated directly from the out- 
put tubes of the amplifier, a 
matching transformer is much 
more satisfactory. It should be 
of the universal type originally 

Fio. 2. Commercial unit. employed for coupling push-pull 

output tubes to variable voice- 
coil taps. As the load on the tube is less than ten milliamperes, 
almost any size of transformer will serve. 

This instrument, properly handled, enables one to inspect the 
output of the sound system, and facilitates the precise setting of the 
exciter lamps, a procedure that is not too well understood by a large 
percentage of projectionists. If the setting is incorrect, jagged 
lines appear upon the mirror when the projector is running without 
film, due to the vibration of the exciter lamp filament, which modu- 
lates the light-beam and produces a characteristic noise oscillogram. 
Proper setting of the lamp smooths this out and assures the best 
possible results. 




Mar., 1937] A NEON-TUBE OSCILLOSCOPE 307 

Tube microphonics show up instantly as jagged lines in the oscillo- 
gram when the tube is touched or gently vibrated. If slight noise is 
indicated, and the tubes and other equipment are apparently in good 
condition, one may suspect a noisy photoelectric cell, which con- 
clusion may be immediately checked by replacing the cell. 

Loose elements in the optical system are more common than is 
suspected. They may be checked by opening the gain control or the 
fader wide, and tapping the various elements with a hammer made of 
a rubber eraser attached to a stick. Poor joints or connections in 
wiring circuits may also be dis- 
covered in the same way, and 
are usually identified by bright 
and jagged noise patterns in the 
oscillogram. 

The instrument has been found 
extremely helpful in setting up 
the bevel gear that drives the 
micarta gear on the constant- 
speed shaft of the Western Elec- 
tric reproducer. A test prod was 
made by placing a long steel 
needle in a Western Electric 2A 

reproducer connected to the regu- 

FIG. 3. Home-made unit, 
lar sound system amplifier. The 

oscilloscope was connected to the output of the amplifier, and a 
sensitive vibration detector was thus provided. The needle was held 
against the center of the constant-speed shaft and also on the bracket 
of the 7 08 A drive. When the gear was too loose the backlash 
registered upon the oscilloscope mirror; when too tight, the gear 
chatter appeared in the form of a distinctive noise oscillogram. Both 
show up as parasitic wave-forms that follow and in some cases over- 
shadow the wave-form registered for proper gear operation. 

The test prod will find many such uses in the projection room 
where slight adjustments and closer checking will effect great im- 
provement in reducing both gear noise and wear. 

A contact type of microphone or pick-up may be substituted for 
the Western Electric 2A, for studying floor and structural vibration 
in projection rooms. Frequently a small wedge properly inserted 
beneath the projector or the motor-generator will effect wonders in 
quieting the operation of the machines. 




308 F. H. RICHARDSON AND T. P. HOVER 

DISCUSSION 

MR. HOVER: 1 believe that the present system of servicing leaves much to be 
desired. It is not entirely the fault of the companies furnishing the service; 
there are union regulations that prevent it from being as successful as it might 
be. That is the reason why I have personally advocated that the projectionist 
himself do his part. 

The average projectionist is a bit "touchy." On the other hand, he has a cer- 
tain pride in his work, and within him feels a certain amount of resentment. I 
believe that resentment was aroused when sound first came in. The attitude of 
the service men who first went into the field was responsible for it. The attitude 
was "I am the authority." They ignored the common courtesies to the men with 
whom they worked. The situation has changed, we all know, but the feeling still 
persists, and the average projectionist is just a bit doubtful about the service 
men. 

Regardless of the amount of equipment that the engineer may bring in, the 
projectionist would rather use a dollar meter or its equivalent for testing, and 




FIG. 4. Mirror and motor assembly. 

use his own brains in applying the meter, than to use one hundred dollars worth 
of equipment that the engineer brings with him. That is the spirit that exists. 

While I am not in a position to state facts and figures, I feel that servicing is 
not covering the theaters that need it. The ones that need it are the little theaters, 
the "shooting galleries," as we call them. After all, they support the industry to 
a large extent. Recently I heard it said that considerable original thought, so far 
as projection is concerned, originates in those theaters because the men are on 
their own; if they can not make inferior equipment work, then the results suffer. 
If developing a little idea such as this instrument will help them, I believe it 
deserves the attention of the engineers. After all, the projectionist has con- 
siderable to say as to whether the picture is going to be a failure or a success. 

MR. WOLF: Then I take it, Mr. Hover, that you advocate that each projection- 
ist service his own equipment? 

MR. HOVER: Not at all. The projectionist should get all the help he can from 
the service men. But, I do not believe that he should be permitted to lean upon 
the service men and depend upon them for the thought and initiative that he 
requires for his own job. 



NEW MOTION PICTURE APPARATUS 



During the Conventions of the Society, symposiums on new motion picture appara- 
tus are held, in which various manufacturers of equipment describe and demonstrate 
their new products and developments. Some of this equipment is described in the fol- 
lowing pages; the remainder will be published in subsequent issues of the Journal. 

NEW RECORDING EQUIPMENT* 

D. CANADY** 



When designing sound-film recording equipment for use in foreign countries, 
for small studios and free-lance cameramen, the limitations imposed by the in- 
tended use of the equipment must be taken into account. Power lines of various 
voltages and frequencies, the lack of skilled technicians, and laboratory facilities 
for processing sound-film are factors that must be carefully considered. 

Much has been said and written during the past few years regarding the rela- 
tive merits of the various methods of recording sound upon film. The methods 
can be roughly divided into two classes : the variable light-source and the con- 
stant light-source. The former appears in the glow-lamp system, and the latter 
in the systems employing the ribbon light-valve and the galvanometer. All 
such systems are capable of recording sound of excellent quality. 

Next to selecting the recorder itself, the most important item is the light-modu- 
lating device. Due to its great simplicity and excellent recording qualities, the 
glow-lamp is ideally suited to the small producer because of several reasons: (1) 
it contains no moving parts, and therefore requires no delicate adjustments; 
(2) it is free from harmonic distortion (harmonics introduced in the other systems 
become serious with increasing frequency); (5) with fixed slit dimensions, glow- 
lamps are capable of recording frequencies beyond the audible limit ; (4) from an 
economy standpoint the glow-lamp is preferable to other types since no additional 
apparatus is required to keep it in working order. 

Glow-lamp recording requires that the exposure be confined to the lower por- 
tion of the H&D curve and is known as toe recording. Toe recording possesses 
certain advantages not found in other methods; some of which are mentioned 
below: 

(1) No special developer is required for the sound negative or positive. One 
bath serves for both negative and positive as well as for picture prints. 

* Received September 29, 1936; presented at the Fall, 1936, Meeting at 
Rochester, N. Y. 

** Canady Sound Appliance Co., Cleveland, Ohio. 

309 



310 



NEW MOTION PICTURE APPARATUS [J. S. M. P. E. 



(2) For the same recording level, toe-recording gives approximately 6 db. more 
volume in reproduction than the methods using the straight-line portion of the 
H&D curve. 

(5) Toe recorded negatives can be reproduced as easily as positives. This is an 

important item when only one print is 
to be made, as for re-recording, pro- 
ducing incidental music for titles, etc. 

(4) Toe recording is less affected 
by slight variations of gamma. A 

^ transmission of 45 to 50 per cent is 

sy satisfactory for both negative and 

positive. 

(5) In practice, toe recording has 
the same modulating range and over- 
load point as other methods. 

Theoretically, straight-line record- 
ing is superior for noise reduction, 
but unless up-to-date processing meth- 




FIG. 1. 



Characteristic curve of 
glow-lamp. 

ods are available and the recording 

equipment is operated by skilled technicians, the toe method will give equal 
results. 

Recording Lamps. In recent months marked improvement has been made in re- 
cording lamps. Present lamps operate at comparatively low voltage and are 




B 



A A 




FIG. 2. (A) 1000-cycle oscillogram of 3-element lamp 
with the third electrode disconnected; (B and C) modu- 
lated to the point at which extinction begins ; (D) resistor 
reduced to 0.05 megohm; (E) series resistor shorted, pro- 
ducing bad distortion. 

quite uniform. Sufficient exposure for toe recording is easily obtained with 
lamps operating at 250 volts and approximately 10 ma. They will withstand 
considerable overload without damage or "smoking up" on the end. A patent 



Mar., 1937] 



NEW MOTION PICTURE APPARATUS 



311 



has recently been granted 1 on an improved type of lamp that delivers sufficient 
light for straight-line recording. 

A feature of the present lamp is its suitability for use in noise-reduction cir- 
cuits. Lamps drawing 10 ma. at 250 volts can be reduced to 1 ma. at 200 volts 
before extinction begins. The spectrum of the emitted light matches very 
closely the sensitivity spectrum of positive recording stock. 

It is well known 2 that light produced by a gas-filled lamp varies in intensity 




FIG. 3. The new sound-film recorder. 



according to the current applied to the electrodes. The time-delay is so small 
that oscillations of the order of one millionth of a second can be faithfully fol- 
lowed. 

The intensity is a function of the applied current, as indicated in Fig. 1. In- 
creasing the polarizing voltage from zero, the lamp does not light until a certain 
value, Vi, has been attained, producing an intensity of D\. From that point 
the intensity is a linear function of the current up to a value F 2 , when the rela- 
tion ceases to be linear. 



312 NEW MOTION PICTURE APPARATUS [J. S. M. P. E. 

If the current is then decreased from the value F 2 , a new curve results. The 
descending curve almost coincides with the ascending curve, but remains slightly 
above it. As the descending curve passes the ignition point Vi, the lamp is not 
extinguished at once, but the intensity continues to decrease steadily almost to 
zero. In order to avoid distortion, it is necessary to limit the operation of the 
lamp between the values Vi and Vz. 

Three-Element Recording Lamps It has been stated 3 that by adding a third 
element to the two-element lamp, a wider modulation range can be obtained with- 
out distortion due to the ionization-maintaining current flowing through the 
third electrode at all times. It has been common practice to connect the ad- 
ditional electrode to the high-voltage supply through a resistance of one or two 
megohms. Neglecting the lamp resistance, a 2-megohm resistor in series with a 
300-volt supply allows only 0.00015 ampere as the steady ionizing current. Re- 
cent investigation has shown that when sufficient current flows through the third 
electrode to enable it to function as a maintaining element, the wave-form is 
distorted in proportion to the current flowing through this element. 




FIG. 4. Recording amplifier. 



A 1000-cycle oscillogram of a three-element lamp operating as a two-element 
lamp with the third element disconnected is shown at A in Fig. 2. B and C (Fig. 2) 
are oscillograms of the lamp modulated to the point where extinction begins. 
Connecting the third electrode to the 300-volt polarizing voltage through a 1- 
megohm resistor had no effect upon the wave-forms B and C. When the resistor 
was reduced to 0.05 megohm, the wave D resulted. E shows a badly distorted 
wave caused by completely shorting out the series resistor. It is apparent that 
the introduction of a third electrode introduces distortion by cutting off the 
peaks, and reduces the modulation range of the glow-lamp instead of increasing it. 

Noise-Reduction Unit. High standards set by the industry demand that the 
light falling upon the unexposed film during recording be reduced to a low value 
during periods of no modulation, for which purpose an auxiliary unit has been 
provided. Design considerations indicate the necessity of keeping the light in- 
tensity within limits not exceeding the straight portion of the H&D curve. Cir- 



Mar., 1937] NEW MOTION PICTURE APPARATUS 313 

cuit design prevents the exposure from dropping below fixed limits, and excessive 
volume will not increase the exposure beyond the straight portion of the curve 
due to saturation of control unit. Patent has been applied for on certain im- 
provements incorporated in this unit, and additional details can not be given at 
this time. 

Recording Lamp PolarizingVoltage Supply. In the past, the polarizing voltage 
required by recording lamps has been largely furnished by heavy-duty radio 
B batteries. Difficulties experienced in obtaining replacement batteries in 
foreign countries and out of the way places has led to the development of a small 
rectifier unit supplying a maximum of 500 volts directly from a 6- or 12-volt 
storage battery. Rectifier units for operation on 110-220 volts a-c. are available 
for studio use. The hum level is reduced to a low value by thorough filtering, and 
no trouble has been experienced from this cause. 

New Sound- Film Recorder. A new recorder possessing unusual constancy of 
speed is shown in Fig. 3. The recorder proper is of cast aluminum, and a 
ribbed base of generous proportions minimizes the vibration of the driving motors 
operating on low-frequency power lines. 

Three flywheels, in addition to a non-resonant sprocket filter, maintain the 
film speed constant. A heavy flywheel is located upon the shaft of the recording 
drum, one of slightly smaller dimensions is used upon the shaft of the driving 
motor, and a third upon the drive sprocket shaft. 

An unusually large "wrap" around the recording drum allows full advantage 
to be taken of the stabilizing effect of the large flywheel. Accurately cut gears 
of special pitch are used throughout, and contribute largely to the quietness 
of the machine. By using "silenced" magazines, the recorder can be used in 
close proximity to microphones when occasion arises. Provision has been made 
for using glow-lamps or any other type of modulator. Ample room has been 
provided in the base for meters, receptacles, switches, etc. 

Recording Amplifier. A semi-portable recording amplifier mounted in an oak 
carrying-case having removable metal panels is shown in Fig. 4. Three mixing 
controls with their associated input receptacles are mounted upon the left panel. 
The top panel includes a filament voltmeter, calibrated volume indicator, and 
a milliammeter. The amplifier is located behind the center panel, which is 
instantly removable for replacing tubes, etc. Plate-circuit jacks are provided in 
the amplifier chassis for testing purposes. Output controls are mounted upon 
the panel at the right, and plate-supply batteries are stored in a compartment 
behind the bottom panel. The amplifier furnishes sufficient amplification so 
that low-level microphones may be used direct without the need of pre- 
amplifiers. 

This form of construction possesses several advantages: (1) less trouble is 
likely to develop when all the components are rigidly mounted together in 
one case; (2) less time is consumed in setting up the equipment, and delays often 
caused by bad connections hi plugs and cables are eliminated; (5) less chance of 
hum pick-up due to the short leads and thorough shielding. 

Another amplifier of extremely light weight is housed in an aluminum case. 
Weight has been reduced to a minimum through the use of new light-weight trans- 
formers and condensers. Amplifiers for studio use are furnished for standard 
rack mounting, and operate on batteries or local power supply. 






314 NEW MOTION PICTURE APPARATUS [J. S. M. p. E. 

REFERENCES 

1 U. S. Patent No. 2,038,825. 

2 MOREAU-HANOT, M.: "Photometric des Lumieres Breves ou Variables," I 
Revue d' Optique Theorique et Instrumental (Paris), Chapt. XI, p. 106 (1934). 

1 BRAMAN,V. T. : "Glow-Lamp Sound-on-Film Recording," Electronics, 2 (June, 
1931), No. 6, p. 679. 



A NEW REEL-END ALARM* 
D. CANADY AND V. A. WELMAN** 



As the duties of the projectionist in a modern motion picture theater become 
more numerous and exacting, and as the quantity and intricacy of the equip- 
ment intrusted to his care is continually increased, some positive, unfailing 
means of announcing to him the impending finish of the reel he is showing upon 
the screen becomes absolutely necessary. 

Without some such signal, tests, adjustments, repairs to equipment, or other 
work of immediate necessity must often be neglected in order to devote the last 
few minutes of a nine- or ten-minute reel to watching for the end, or, as an al- 
ternative, doing his other tasks and taking a chance on catching the change-ovef 
when the time comes. 

Even in the silent days, when the operation was simpler, there was a general 
demand for some such signal, the earliest, perhaps, being a coin wound into the 
reel of film fifty or a hundred feet from the end. When the film unrolled to the 
proper point the coin would drop and the resulting clatter would provide the 
signal. 

The most common device, both of the silent days and at the present time, is an 
arm pivoted near the edge of the magazine, bearing upon its free end a roller which 
rests on the film as it unwinds. When the core of film is no longer large enough 
to support it the roller drops with a bang, sounding a warning. 

Thousands of this type of reel-end alarm are in use, and these thousands in 
constant daily use have caused and are causing an immense amount of film 
damage. It is no use to argue that with proper care it need not damage the film, 
or that if properly designed no damage will result. The proof of the matter 
may be had by dropping into any second-run theater in the country and watching 
the wavy scratch down the center of almost every reel. 

Since the roller rides upon the emulsion side of the film, which, in first-run 
houses is usually fairly "green," this type of alarm is particularly dangerous in 
those houses. 

This type of reel-end alarm has the further disadvantage that it must be manu- 

* Received September 21, 1936; presented at the Fall, 1936, Meeting at 
Rochester, N. Y. 
** Canady Sound Appliance Co., Cleveland, Ohio. 



Mar., 1937] 



NEW MOTION PICTURE APPARATUS 



315 



ally reset each time a reel is threaded in the projector. The human factor is 
thereby introduced, which is especially likely to fail during periods of difficulty 
with other parts of the projection room equipment, just at the times when the 
reel-end alarm is of most importance. 

The requirements for a successful reel-end alarm are: 

(2) It should accurately and positively announce the approach of the end of 
the reel, a predetermined distance from the end. 

(2) It should not damage the film in any way. 

(5) It should be automatic in action, requiring no act on the part of the pro- 
jectionist to place it in operation. 




FIG. 1. Schematic arrangement of reel-end alarm. 

(4) It should be rugged, and easy to install and maintain. These require- 
ments are fulfilled in the alarm described here. Experimental samples were con- 
structed and used about five years ago. The patent was applied for in 1932 and 
was recently issued as No. 1,982,133. The principle employed is that of uncover- 
ing a beam of light so that the light will fall upon a light-sensitive cell when the 
film unwinds to a predetermined point, a principle well known now and applied to 
many uses during the past year, but something of a mystery at the time the 
first models of this reel-end alarm were constructed. 

The patent drawing (Figs. 1 and 2) is self explanatory, although in actual ap- 
plication many variations are possible as to light-source, lens, adjustment, type 
of light-sensitive cell, and relay and signalling device. The apparatus consists 
essentially of a light-source and a focusing and adjusting device, in a case at- 
tached to the upper magazine; and a light-sensitive cell to receive the beam of 



316 NEW MOTION PICTURE APPARATUS IJ. S. M. P. E. 

light after it has passed the core of the reel of film. The light-sensitive cell is 
connected to a suitable relay, either vacuum-tube or magnetic, which operates a 
signal consisting of lights, buzzer, or electric bell. A one-stroke bell is perhaps 
the best signal. The power for the circuit is derived from the commercial 
power lines, no batteries being required. 

As can be seen from Fig. 1, when the reel of film has been reduced to a prede- 
termined size, the beam of light is no longer interrupted but proceeds to the light- 
sensitive cell. Perhaps the most compact and most easily installed arrangement 
is to house the light-source, a copper-oxide cell, relay, and bell with their necessary 
associated parts in one metal box, mounted directly upon the magazine, returning 
the beam of light from the exciter lamp to the photoelectric cell stationed along- 




FIG. 2. Photograph of installation. 

side it by means of a glass or metal mirror attached to the magazine upon the 
opposite side of the reel. Such an installation would require a minimum of 
changes in the magazine, the only outside electrical connection being a rubber- or 
metal-covered cord to the framing-light circuit or other convenient lighting cir- 
cuit. 

DISCUSSION 

MR. CRABTREE: What is the sensitivity of the reel-end alarm, in terms of film 
convolutions? 

MR. WELMAN: In this experimental set, two or three layers of film. 

MR. CRABTREE : Can any projectionist say whether that sensitivity is adequate? 

MR. HOVER: I believe that that is more sensitive than the average roller 
alarm. I checked the roller type some time ago, and found that it will drop, as 
a rule, anywhere within a length of film of approximately 12 feet one way or the 



Mar., 1937] NEW MOTION PICTURE APPARATUS 317 

other. Even the most careful and accurate set-up has a certain amount of play 
in the bracket or arm that holds the rolls, and it is also influenced by whether the 
film is wound tightly or loosely. 

MR. CRABTREE: When the usual black spot appears upon the screen, what 
tolerance does the projectionist have? In other words, how many seconds may 
elapse before he throws the change-over switch? 

MR. WELMAN: This appliance has nothing to do with change-over. This pro- 
vides a signal indicating that the change-over dot is soon coming. It is to warn 
the projectionist to look for the dot. 

MR. CRABTREE: Why could it not be used to eliminate the black spot? 

MR. WELMAN: It is not accurate enough for that. To eliminate the dot you 
must hit it within a couple of frames. 

MR. TOWNSEND: I used to disapprove reel-end alarms for the reason that I 
felt it gave the projectionist a chance to nap, or do something other than pay at- 
tention to the show. Since I have gone back into the projection room I have used 
the alarm myself, not so that I can take a nap, but because the alarm has certain 
advantages, especially in smaller theaters that are now using the Suprex carbons 
which burn at quite a high rate. 

The reel-end alarm provides a very definite time in which to light the arc. The 
projectionist can gauge the length of carbon, and can use short lengths to ad- 
vantage. The human factor enters in this way: A projectionist who happens to 
be of a nervous temperament is likely to strike the arc two or three minutes earlier 
than is necessary if he merely looks at the amount of film left upon the reel. If 
he has an alarm that is thoroughly accurate, he will not light up too soon. 
Another projectionist may strike the arc too late, so as to utilize as much of the 
carbon as possible. He may change over and find that he has a poor light 
because of having struck the arc too late. 

MR. HOVER : A signal of this type would be greatly appreciated by projection- 
ists who have gone through a fire in the projection room. After such an ex- 
perience the projectionist will have a strong aversion to opening the magazine 
door when the machine is running. 

The condition actually exists, that because the little ports, or windows, in the 
magazines are not well lighted, the magazine door is sometimes left open so that 
the projectionist can watch the film. Every projector should have some kind 
of device that will indicate the time to light up the arc without having to open the 
magazine door. 

MR. DANIELSON: Is the danger of fire increased by putting an exciter lamp in 
the upper magazine? What do the fire authorities think of the device? 

MR. WELMAN: The lamp is in a separate box, and is blocked off by a glass 
plate. It is no worse than the light now in the magazines, which the under- 
writers have approved. 



A DEMONSTRATION TRIODE FOR VISUALIZING ELECTRONIC 
PHENOMENA* 



F. E. ELDREDGE AND H. F. DART** 



Several discoveries that were later to prove fundamentally important in the 
design of electronic tubes were made shortly before the turn of the twentieth 
century. Edison, in 1889, had just invented a practical and commercial electric 
lamp. In 1895, Roentgen, the German, showed his first x-ray pictures. In 1896, 
Becquerel, the Frenchman, discovered radioactivity. Two years later, radium 
was discovered by the Curies in France. From these discoveries arose many new 
concepts of physics and many developments that later evolved into extremely 
useful tools. 

Starting in 1900, many theories of physics required an electron. Milliken mea- 
sured the electric charge of the electron and found the value so minute that it is 
difficult to picture its size. In their book, Photocells and Their Applications. 
Zworkin and Wilson say: 

"To realize how small the mass of an electron is, imagine an electron speeding 
through space at the rate of 100 miles an hour or about 150 feet per second. Let 
us calculate the force of a blow it would give to an obstacle placed in its path, 
assuming the electron was brought to a dead stop in one-tenth second. We will 
find the average force experienced by the obstacle would have the amazingly 
small value of one hundred-thousand-million-billion-trillionth of a pound, or 
l/100,000,000,000,000,000,000,000,000,0001b." 

It is interesting to note how quickly the fields of industry and amusement have 
put the new ideas into practical, every-day use. In fact, one of the amazing char- 
acteristics of this age is not only the willingness to accept changes, but the will to 
utilize the achievements of science and scientific minds. 

To be more specific, let us consider the filament of a lamp. It receives heat 
energy, which causes luminescence. An incidental by-product is the liberation of 
electrons near this filament. In a lamp, however, they do not serve any direct 
useful purpose. Placing to one side of a heated filament a piece of metal or a plate, 
on which a positive voltage is applied, will produce an appreciable flow of current 
between the heated filament and the plate. A modification of this simple device 
has considerable importance, for in a proper circuit it will convert alternating 
into direct current suitable for operating arcs of motion picture equipment, for 

* Received September 17, 1936; presented at the Fall, 1933, Meeting at 
Rochester, N. Y. 

** Special Products Sales Department, Westinghouse Lamp Co., Bloomfield, 
N. J. 
318 



NEW MOTION PICTURE APPARATUS 



319 



charging storage batteries, for driving trolley cars, and for a host of other 

similar applications. 

Some thirty years ago, Dr. Lee de Forest discovered that a grid, placed be- 
tween the heated filament and the plate of 
the tube, exerted a definite control over the 
passage of electrons from the filament to the 
plate. The result was the three-electrode 
tube, which depends for its operation upon 
the unidirectional flow of electrons from a 



JT 





FIG. 2. (Above) Circuit for oper- 
ating the WL-787 triode on direct 
current. 

FIG. 3. (Below) Circuit for oper- 
ating the WL-787 triode on alternating 
current. 

thermionic cathode to a positive plate, with 
a grid between the two. A small voltage 
applied to this grid can oppose or add to the 
effect of voltage on the plate. In other 
words, the grid permits controlling the space 
current flowing between the plate and the 
cathode. 

The action of the grid may be likened to 
that of a valve in a water-pipe. A small 
amount of energy, applied to opening and closing the valve, controls a much 
larger amount, represented by the water under pressure. The grid is usually 
operated at a potential negative with respect to the cathode, since then no current 
is drawn by it, and, consequently, no power, 



FIG. 1. The effect of the elec- 
trons striking the plate, which is 
coated with willemite, is seen in 
luminous bands. The width of 
these light bands may be varied 
by changing the voltage of the 
grid or the plate. 



320 NEW MOTION PICTURE APPARATUS [J. S. M. P. E. 

Thus, within a few years a new tool the electronic tube has been developed, 
which has since become indispensable in the fields of communication, manu- 
facturing, amusement, medicine, and many others. Typical examples are the 
universal use of electronic tubes in radio transmitting and receiving equipment; 
long-distance telephone lines; the quick response a decade ago of the motion 
picture industry to add sound to their films, thereby utilizing phototubes, recti- 
fiers, and amplifier tubes; oscillator tubes in short-wave therapy equipment; 
and now industries of all kinds are using phototubes, amplifiers, high- and low- 
voltage rectifiers, and grid-glow tubes in the initiation and control of many manu- 




FIG. 4. Showing the magnetic effect upon the fluorescent 
pattern. 

facturing processes. There is now a degree of reliability and constancy of char- 
acteristics not approached by the tubes of a decade ago. 

"SEEING" ELECTRONS BY FLUORESCENCE 

The widespread use of electronic tubes in all fields of endeavor has led to the 
introduction of courses of instruction in electronics in the curricula of many 
universities, colleges, and technical schools. The Westinghouse Demonstration 
Triode WL-787 was developed for use in studying the fundamental principles and 
operation of the three-electrode tube. It shows visually, in a manner impossible 
to accomplish in any other way, exactly what takes place when changes are made 
in the grid and plate voltages of a vacuum tube. 

By varying the grid voltage in steps, the effect of changing excitation upon the 
electron flow from the filament to the plate, and the correlation of this action 
with fluorescence on the anode, can readily be shown. The controlling effect of 
the grid upon the electrons radiated by the filament is demonstrated by the 
fluorescent pattern upon the plate. When greater and greater negative voltage 



Mar., 1937] 



NEW MOTION PICTURE APPARATUS 



321 



is applied to the grid, the fluorescent strips opposite the openings of the grid 
become narrower and narrower, until finally they disappear. This condition cor- 
responds to the grid-bias value at which plate current cut-off occurs, and tinder 
which condition no electrons will reach the plate. 

The grid may be made still more negative, without further apparent change. 
Such condition corresponds to that existing 
in an actual triode when the net voltage 
of the grid is greater than that required to 
produce plate-current cut-off. 

Lessening the negative voltage of the grid 
will cause the fluorescence to reappear, and 
the width of the strips will increase as the 
net voltage is decreased. The fluorescence 
will be proportional to the rate at which the 
electrons reach the plate, which is also a 
measure of the plate current. As the grid 
becomes positive with respect to the fila- 
ment, the fluorescent lines become still wider 
and wider; until, when the grid becomes 
positive, the fluorescence covers the entire 
plate with quite uniform intensity. 




GRID VOLTAGE 

FIG. 5. Average grid character- 
istic of the WL-787 triode; charac- 
teristics of individual tubes may 
vary widely from those shown. 



EFFECT OF DISTORTION 

Distortion in an amplifier may be demon- 
strated by using a large excitation signal 
or a wide range of grid voltage. It may 
thus be shown that the width of the fluores- 
cent bands does not vary in proportion to changes of grid voltage, particularly 
when the excitation is so great as to cause plate-current cut-off. In the 
latter case the width of the fluorescent bands will not accurately follow the 

excitation voltage values, corresponding to 
the conditions existing in a triode when 
there is distortion in the ouput. 

The amplifying effect of the tube may 
be demonstrated by noting the grid and 
plate voltages for a certain amount of 
fluorescence on the plate, and then chang- 
ing the grid voltage by a few volts, say, 
10. It will be found that the plate 
voltage must be changed by a much 
greater amount to restore the fluorescent 
bands to their original width. The ratio 
of the change of voltage of the plate 
required to compensate for the change of 
grid voltage is the amplification factor of the tube. 

Another interesting demonstration is to hold a strong magnet near the side of 
the plate. In addition to showing visually the effect of the magnetic field upon 




FIG. 6. Average plate character- 
istic of the WL-787 triode. 



322 



NEW MOTION PICTURE APPARATUS 



the electrons and their distribution, it is possible under favorable conditions to 
obtain a representation of the lines of magnetic force. 

The filament consists of several parallel oxide-coated wires, all of which are 
located in one plane, so as to distribute the plate current uniformly. The anode 
is the usual flat plate mounted parallel to the plane of the filament. The grid is 
a fairly open and conventional structure mounted between the filament and the 
plate. A coating of willemite on the side of the plate toward the grid and filament 
shows a bright greenish fluorescence when bombarded by the electrons emitted 
by the filament. The glow is pronounced and clearly visible wherever the elec- 
trons strike, producing a definite pattern of the grid upon the plate. 

TABLE I 



Filament 

Plate 

Grid 



Design Data and Ratings of Type WL-787 Triode 



Voltage 
Normal Max. 



6.0 



6.3 
300 

:200 



Normal 

1.6 a. 



Current 



Max. 

100 ma. 
50 ma. 



Amplification factor 

Filament 

Plate size 

Overall height 

Maximum diameter 

Base 

Socket 



2 (approx.) 
Oxide-coated 

3 X I 1 / 2 inches (approx.) 
10 inches 

2 5 /s inches 
4-pin industrial 
Ser. No. 793202 



The size and arrangement of the parts have been made such as to render the 
tube useful in demonstrating the action of the grid in a three-electrode tube. 
The plate is large enough to permit the action to be visible to everyone in a lecture 
or classroom of reasonable size. A slight amount of experimentation will show 
how to handle the tube to demonstrate the desired effects. The tube is practically 
fool-proof, and will withstand a wide variety of operating conditions. Either 
alternating current or direct current may be used to heat the filament and to sup- 
ply voltages for the grid and plate. A tube socket and adjustable sources of volt- 
age are all that are needed to operate the tube, although a few meters will be con- 
venient for making adjustments and readings. 



COMMITTEES 

of the 
SOCIETY OF MOTION PICTURE ENGINEERS 

(Correct to February 20th; additional appointments or changes may be made at 
any time during the year as necessity or expediency may require.) 



L. W. DAVEE 
H. GRIFFIN 



ADMISSIONS AND TRANSFERS 

G. FRIEDL, JR., Chairman 
S. HARRIS 
D. E. HYNDMAN 



P. J. LARSEN 
M. W. PALMER 



A. N. GOLDSMITH 
A. C. HARDY 



BOARD OF EDITORS 

J. I. CRABTREE, Chairman 
L. A. JONES 
E. W. KELLOGG 



H. G. KNOX 
T. E. SHEA 



COLLEGE COURSE IN TECHNICAL MOTION PICTURE EDUCATION 
T. E. SHEA, Chairman 



A. N. GOLDSMITH 



L. A. JONES 



W. H. CARSON 
O. O. CECCARINI 



COLOR 

J. A. BALL, Chairman 
C. H. DUNNING 
R. M. EVANS 



A. M. GUNDELFINGER 

H. W. MOYSE 



CONVENTION 

W. C. KUNZMANN, Chairman 
H. GRIFFIN P. MOLE K. F. MORGAN 

J. H. KURLANDER M. W. PALMER 

EUROPEAN ADVISORY COMMITTEE 

J. VAN BREUKELEN, Chairman 
H. WARNCKE F. H. HOTCHKISS I. D. WRATTEN 



EXCHANGE PRACTICE 

A. W. SCHWALBERG, Chairman 

O. C. BINDER K. C. KAUFMAN H. RUBIN 

A. S. DICKINSON J. S. MACLEOD J. P. SKELLY 

G. K. HADDOW H. A. MERSAY J. H. SPRAY 

N. F. OAKLEY 



323 



324 



COMMITTEES OF THE SOCIETY 



[J. S. M. P. E. 



J. E. ABBOTT 
T. ARMAT 



O. B. DEPUE 



A. N. GOLDSMITH 



HISTORICAL 

E. THEISEN, Chairman 
G. A. CHAMBERS 
W. CLARK 

MUSEUM 

(Western) 

E. THEISEN, Chairman 
J. A. DUBRAY 

HONORARY MEMBERSHIP 
J. I. CRABTREE, Chairman 



G. E. MATTHEWS 
T. RAMSAYE 



A. REEVES 



E. A. WILLIFORD 



A. C. HARDY 



JOURNAL AWARD 

E. A. WILLIFORD, Chairman 
E. HUSE 



G. F. RAGKETT 



L. A. BONN 
R. M. EVANS 
G. GIBSON 
E. HUSE 



D. P. BEAN 
F. E. CARLSON 
W. B. COOK 



LABORATORY PRACTICE 

D. E. HYNDMAN, Chairman 
T. M. INGMAN 
C. L. LOOTENS 
R. F. MITCHELL 



H. W. MOYSE 

J. M. NlCKOLAUS 

W. A. SCHMIDT 
J. H. SPRAY 



NON-THEATRICAL EQUIPMENT 

R. F. MITCHELL, Chairman 

H. A. DEVRY . J. H. KURLANDER 

E. C. FRITTS A. SHAPIRO 

H. GRIFFIN A. F. VICTOR 

J. A. HAMMOND 



PAPERS 

G. E. MATTHEWS, Chairman 

C. N. BATSEL M. E. GILLETTE T. E. SHEA 

L. N. BUSCH E. W. KELLOGG P. R. VON SCHROTT 

A. A. COOK R. F. MITCHELL H. C. SILENT 

L. J. J. DIDIEE W. A. MUELLER I. D. WRATTEN 

H. B. SANTEE 



J. E. ABBOTT 
J. I. CRABTREE 
A. S. DICKINSON 



PRESERVATION OF FILM 

J. G. BRADLEY, Chairman 
R. EVANS 
M. E. GILLETTE 
C. L. GREGORY 



T. RAMSAYE 
V. B. SEASE 
W. A. SCHMIDT 



Mar., 1937] 



COMMITTEES OF THE SOCIETY 



325 



MEMBERSHIP AND SUBSCRIPTION 



Alabama 


Michigan 


Ohio 


P. A. KING 


]. F. STRICKLER 


C. C. DASH 






R. H. GILES 


California 


Minnesota 


V. C. WELMAN 


J. O. AALBERG 


C. L. GREENE 




C. W. HANDLE Y 


R. H. RAY 


Pennsylvania 


E. HUSE 




H. BLOOMBERG 


R. H. McCULLOUGH 


Missouri 


A. GOODMAN 


G. A. MITCHELL 


J. S. COPLEY 


I. SAMUELS 


P. MOLE 






K. F. MORGAN 


New York 


Texas 


W. A. MUELLER 


A. BECKER 


H. H. FRASCH 


H. G. TASKER 


F. E. CAHILL 






A. A. COOK 


District of Columbia 


Georgia 


A. S. DICKINSON 


H. T. COWLING 


N. WEIL 


J. J. FINN 


R. EVANS 




J. FRANK, JR. 


N. D. GOLDEN 


Illinois 


S. HARRIS 


F. J. STORTY 


H. A. DEVRY 


D. E. HYNDMAN 




B. J. KLEERUP 


W. H. INGRAM 


Travelling 


S. A. LUKES 


O. E. MILLER 


E. AUGER 


C. G. OLLINGER 


F. H. RICHARDSON 


C. BRENKERT 


J. M. SCHAEFER 


P. D. RIES 


F. HOHMEISTER 


J. H. TOLER 


C. J. STAUD 


W. C. KUNZMANN 




L. M. TOWNSEND 


D. McRAE 


Massachusetts 


J. S. WARD 


O. F. NEU 


J. S. ClFRE 




H. H. STRONG 


S. SUMNER 


Japan 




A. B. WEST 


T. NAGASE 


Germany 




Y. OWAWA 


W. F. BIELICKE 


Australia 






H. C. PARRISH 


New Zealand 


Hawaii 




C. BANKS 


L. LA CHAPELLE 


Austria 






P. R. VON SCHROTT 


England 


Holland 




W. F. GARLING 


J. VAN BREUKELEN 


Canada 


R. G. LlNDERMAN 




F. C. BADGLEY 


E. McM ASTER 


India 


G. H. BATTLE 


R. TERRANEAU 


G. D. LAL 


C. A. DENTELBECK 




H. S. MEHTA 


B. E. NORRISH 


France 


M. L. MISTRY 




L. J. DIDIEE 




China 


L. G. EGROT 


Russia 


R. E. O'BOLGER 


F. H. HOTCHKISS 


E. G. JACHONTOW 



326 



COMMITTEES OF THE SOCIETY 



[J. S. M. p. E. 



L. N. BUSCH 
G. A. CHAMBERS 
A. A. COOK 



PROGRESS 

J. G. FRAYNE, Chairman 
R. M. CORBIN 
R. E. FARNHAM 
W. LEAHY 



G. E. MATTHEWS 
V. E. MILLER 
G. WORRALL 



J. I. CRABTREE 



J. O. BAKER 
T. C. BARROWS 
F. E. CAHILL 
J. R. CAMERON 
A. A. COOK 
J. K. ELDERKIN 
J. J. FINN 
R. R. FRENCH 
E. R. GEIB 



PROGRESS AWARD 

A. N. GOLDSMITH, Chairman 

M. C. BATSEL R. M. EVANS 



PROJECTION PRACTICE 

H. RUBIN, Chairman 
A. N. GOLDSMITH 
A. GOODMAN 
H. GRIFFIN 
S. HARRIS 
J. J. HOPKINS 

C. F. HORSTMAN 

D. E. HYNDMAN 
P. A. McGuiRE 

R. MlEHLING 



E. R. MORIN 
M. D. O'BRIEN 
G. F. RACKETT 

F. H. RICHARDSON 

B. SCHLANGER 

C. TUTTLE 

J. S. WARD 
V. A. WELMAN 
A. T. WILLIAMS 



J. R. CAMERON 
J. J. FINN 



PUBLICITY 

W. WHITMORE, Chairman 
S. HARRIS 
G. E. MATTHEWS 
W. A. MUELLER 



P. A. McGuiRE 
F. H. RICHARDSON 



P. ARNOLD 
M. C. BATSEL 
F. C. BADGLEY 
L. N. BUSCH 
A. CHORINE 
A. COTTET 
L. DE FEO 
A. C. DOWNES 

J. A. DUBRAY 

P. H. EVANS 



STANDARDS 

E. K. CARVER, Chairman 
R. E. FARNHAM 
C. L. FARRAND 
G. FRIEDL, JR. 
H. GRIFFIN 
A. C. HARDY 
R. C. HUBBARD 
E. HUSE 
C. L. LOOTENS 
K. F. MORGAN 
T. NAGASE 



N. F. OAKLEY 
G. F. RACKETT 
W. B. RAYTON 
C. N. REIFSTECK 
H. RUBIN 

0. SANDVIK 
H. B. SANTEE 
J. L. SPENCE 

J. VAN BREUKELEN 

1. D. WRATTEN 



STUDIO LIGHTING 

R. E. FARNHAM, Chairman 
W. C. KUNZMANN V. E. MILLER E. C. RICHARDSON 

J. H. KURLANDER M. W. PALMER F. WALLER 

G. F. RACKETT 



[ar., 1937] COMMITTEES OF THE SOCIETY 327 

SECTIONS OF THE SOCIETY 

(Atlantic Coast) 

G. FRIEDL, JR., Chairman 

L. W. DAVEE, Past-Chairman M. C. BATSEL, Manager 

D. E. HYNDMAN, Sec.-Treas. H. GRIFFIN, Manager 

(Mid-West) 

C. H. STONE, Chairman 

R. F. MITCHELL, Past-Chairman O. B. DEPUE, Manager 

S. A. LUKES, Sec.-Treas. B. E. STECHBART, Manager 

(Pacific Coast) 

K. F. MORGAN, Chairman 

H. G. TASKER, Past- Chairman J. O. AALBERG, Manager 

G. A. CHAMBERS, Sec.-Treas. H. W. MOYSE, Manager 



SPRING, 1937, CONVENTION 
SOCIETY OF MOTION PICTURE ENGINEERS 

HOLLYWOOD-ROOSEVELT HOTEL 
HOLLYWOOD, CALIF. 
MAY 24th-28th, INCL. 

Officers and Committees in Charge 

W. C. KUNZMANN, Convention Vice-President 

J. I. CRABTREE, Editorial Vice-President 

H. G. TASKER, Past-President 

G. F. RACKETT, Executive Vice-President 

K. F. MORGAN, Chairman, Pacific Coast Section 

G. E. MATTHEWS, Chairman, Papers Committee 

Local Arrangements and Reception Committee 

P. MOLE, Chairman 

E. HUSE K. F. MORGAN H. G. TASKER 

J. O. AALBERG C. W. HANDLEY G. F. RACKETT 

C. N. BATSEL R. H. McCuLLOUGH H. W. MOYSE 

G. A. CHAMBERS G. S. MITCHELL W. J. QTJINLAN 

Information and Registration 

W. C. KUNZMANN, Chairman 
E. R. GEIB S. HARRIS C. W. HANDLEY 

Ladies 1 Reception Committee 

MRS. K. F. MORGAN and MRS. P. MOLE, Hostesses 

assisted by 

MRS. F. C. COATES MRS. E. HUSE MRS. E. C. RICHARDSON 

MRS. C. W. HANDLEY MRS. W. J. QUINLAN MRS. H. G. TASKER 
MRS. G. F. RACKETT 

Banquet 

E. HUSE, Chairman 

W. C. KUNZMANN K. F. MORGAN G. F. RACKETT 

P. MOLE W. J. QUINLAN H. G. TASKER 

328 



SPRING CONVENTION 329 

Projection Committee 

H. GRIFFIN, Chairman 

J. O. AALBERG J. FRANK, JR. C. W. HANDLEY 

L. E. CLARK J. G. FRAYNE R. H. McCuLLOUOH 

G. M. GROSJBAN 
Officers and Members of Los Angeles Local No. 150, I.A.T.S.E. 

Hotel Accommodations Committee 

G. F. RACKETT, Chairman 
E. HUSE K. F. MORGAN 

W. C. KUNZMANN H. G. TASKER 

H. C. SILENT 

Transportation Committee 

C. W. HANDLEY, Chairman 
G. A. CHAMBERS S. HARRIS 

H. GRIFFIN F. E. JAMES 

Publicity 

W. WHITMORE, Cfiairman 
J. J. FINN S. HARRIS 

W. GREENE G. E. MATTHEWS 

W. A. MUELLER 

Membership 

E. R. GEIB, Chairman 
G. A. CHAMBERS W. GREENE S. HARRIS 

Headquarters 

Headquarters of the Convention will be the Hollywood-Roosevelt Hotel, where 
excellent accommodations are assured. A reception suite will be provided for the 
Ladies' Committee, and an excellent program of entertainment will be arranged 
for the ladies who attend the Convention. 

Special hotel rates, guaranteed to SMPE delegates, European plan, will be as 
follows : 

One person, room and bath $ 3 . 50 

Two persons, double bed and bath 5.00 
Two persons, twin beds and bath 6 . 00 
Parlor suite and bath, 1 person 9 . 00 

Parlor suite and bath, 2 persons 12.00 

Room reservation cards will be mailed to the membership of the Society in the 
near future, and everyone who plans to attend the Convention should return his 
card to the Hotel promptly in order to be assured of satisfactory accommodations. 
Special garage rates will be provided, for SMPE delegates who motor to the 
Convention. 






330 SPRING CONVENTION [J. S M. P. E. 

Railroad Fares 

The dates of the Convention have been chosen in order that delegates may avail 
themselves of the summer tourists' rates, which go into effect May 15th. The 
following table lists the railroad fares and Pullman charges: 

Railroad 

Fare Pullman 

City (round trip) (one way) 

Washington $120.75 $20.50 

Chicago 86.00 15.75 

Boston 132.80 22.25 

Detroit 98.30 18.00 

New York 126.90 21.75 

Rochester 112.50 19.25 

Cleveland 101.35 18.00 

Philadelphia 122.85 21.00 

Pittsburgh 107.10 18.75 

The railroad fares given above are for round trips, forty-five day limits. Ar- 
rangements may be made with the railroads to take different routes going and 
coming, if so desired, but once the choice is made it must be adhered to, as changes 
in the itinerary may be effected only with considerable difficulty and formality. 
Delegates should consult their local passenger agents as to schedules, rates, and 
stop-over privileges. 

New streamlined trains will be operating from Chicago to Los Angeles and San 
Francisco, making the trip to Los Angeles in 39 hours. Special fares are levied on 
these trains. 

Technical Sessions 

An attractive program of technical papers and presentations is being arranged 
by the Papers Committee. Several sessions will be held in the evening, to permit 
those to attend who would be otherwise engaged in the daytime. All sessions will 
be held at the Hotel. 

Semi- Annual Banquet 

The Semi-Annual Banquet of the Society will be held at the Hotel on Wednes- 
day, May 26th. Addresses will be delivered by prominent members of the indus- 
try, followed by dancing and entertainment. Tables reserved for 8, 10, or 12 per- 
sons; tickets obtainable at the registration desk. 

Inspection Tours and Diversions 

Arrangements are under way to visit one or more of the prominent Hollywood 
studios, and passes will be available to registered members to several Hollywood 
motion picture theaters. Arrangements may be made for golfing and for special 
trips to points of interest in and about Hollywood. 



Mar., 1937] SPRING CONVENTION 331 

Ladies' Program 

An especially attractive program for the ladies attending the Convention is be- 
ing arranged by Mrs. K. F. Morgan and Mrs. P. Mole, hostesses, and their Ladies' 
Committee. A suite will be provided in the Hotel, where the ladies will register 
and meet for the various events upon their program. Further details will be pub- 
lished in a succeeding issue of the JOURNAL. 

Points of Interest 

En route: Boulder Dam, Las Vegas, Nevada; and the various National Parks. 
Hollywood and vicinity: Beautiful Catalina Island; Zeiss Planetarium; Mt. 
Wilson Observatory; Lookout Point, on Lookout Mountain; Huntington Li- 
brary and Art Gallery (by appointment only); Palm Springs, Calif.; Beaches at 
Ocean Park and Venice, Calif.; famous old Spanish missions; Los Angeles Mu- 
seum (housing the SMPE motion picture exhibit); Mexican village and street, 
Los Angeles. 

TENTATIVE PROGRAM 
Monday, May 24th 
10:00 a.m. Blossom Room 
Registration 
Society Business 
Committee Reports 
Technical Papers Program 
12:30 p.m. Florentine Room 

Informal Get-Together Luncheon for members, their families, and 
guests. Brief addresses by several prominent members of the 
industry. 
2:00 p.m. Blossom Room 

Technical Papers Program. 
8:00 p.m. (To be announced later.) 

Tuesday, May 25th 

10:00 a.m. Blossom Room 

Technical Papers Program 
2:00 p.m. (To be announced later.) 
8:00 p.m. Blossom Room 

Technical Papers Program 

Wednesday, May 26th 

10:00 a.m. Blossom Room 

Technical Papers Program 
2:00 p.m. (To be announced later.) 
8:00 p.m. Blossom Room 

Semi- Annual Banquet and Dance of the SMPE; Addresses by 
eminent members of the industry; dancing and entertainment. 



332 SPRING CONVENTION 

Thursday, May 27th 

10:00 a.m. Open morning 
2:00 p.m. Blossom Room 

Technical Papers Program 
8:00 p.m. Blossom Room 

Technical Papers Program 

Friday, May 28th 

10:00 a.m. Blossom Room 

Technical Papers Program 
2:00 p.m. Blossom Room 

Technical Papers Program 

Open Forum 

Adjournment of the Convention 

NOTE: All technical sessions will be held in the Blossom 
Room of the Hollywood-Roosevelt Hotel. There will be n 
public exhibit of apparatus anywhere in the Hotel; although 
members registered in the Hotel will, of course, be privileged 
to display any equipment they wish in their own rooms. 






SOCIETY ANNOUNCEMENTS 



PACIFIC COAST SECTION 

At a meeting held on January 28th at the auditorium of the Hollywood Cham- 
ber of Commerce, a paper by Dr. C. E. K. Mees, Vice- President and Director of 
Research, Kodak Research Laboratories, Rochester, N. Y., was presented, en- 
titled, "Interesting Phases in the Historical Development of Photography, 
Including Recent Progress." As the subject was of wide interest to the industry, 
the meeting was well attended. 

Arrangements are being made by the Board of Managers under the Chair- 
manship of Mr. K. F. Morgan to conduct a series of interesting meetings each 
month during the season. Members of the Pacific Coast Section are reminded 
that the Society has recently established a West Coast office in the Taft Building, 
at Hollywood Boulevard and Vine St., Hollywood, under the Management of 
Mr. Walter R. Greene. 

MID-WEST SECTION 

The regular monthly meeting of the Section was held on February llth in the 
review room of Paramount Pictures Distributing Corp., Chicago, 111. Mr. E. R. 
Geib, of National Carbon Company, Cleveland, Ohio, presented "The Story of the 
Carbon Arc," a discourse and demonstration film showing the process of carbon 
manufacture and the evolution of the carbon arc. 

The next meeting of the Section has been scheduled for Thursday, March llth. 



ADMISSIONS COMMITTEE 

At a recent meeting of the Admissions Committee, at the General Office of the 
Society, the following applicants for membership were admitted to the Associate 
grade: 



COMI, P. E. DENNISON, L. I. 

Theater Service & Supply Co. Patterson & Dennison, Inc. 

112 Arlington St. 6210 Greenfield Ave. 

Boston, Mass. West Allis, Wis. 

CROWE, J. W., JR. DICKERSON, A. B. 

837 W. 36th Place 738 Woodward Building 

Los Angeles, Calif. Washington, D. C. 

HADDOW, G. K. 

DALEY, T. S. Paramount Pictures, Inc. 

Imperial Theater 1501 Broadway 

Toronto, Ontario, Canada New York, N. Y. 

333 



334 



SOCIETY ANNOUNCEMENTS 



HESTERMAN, W. H. 
6137 Addison St. 
Chicago, 111. 

JOHNSON, G. A. 

Eastman Kodak Co. 
Rochester, N. Y. 

KAUFMAN, H. C. 

Columbia Pictures Corp. 
729 Seventh Ave. 
New York, N. Y. 

LAMSON, F. M. 
2901 First Ave., South 
Seattle, Washington 

LEROY, F. S. 

RCA Manufacturing Co., Inc. 
Camden, N. J. 

RAPAROWITZ, G. A. 
1505 Wicker Park Ave. 
Chicago, 111. 



RATHMELL, R. 

RCA Victor Co. of China 
P. O. Box 123 
Hongkong, China 

REINHOLD, L. 
95 Hillcrest St. 
Great Kills 
Staten Island, N. Y. 

REYNOLDS, D. J. 
272 Deverill St. 
Ludlow, Ky. 

SCHEFFY, H. 

1279 Lake Ave. 
Rochester, N. Y. 

SCHOENBERG, M. 

1540 Broadway 
New York, N. Y. 

SKELLY, J. P. 

1514 Albany Ave. 
Brooklyn, N. Y. 






In addition, the following applicants have been admitted by vote of the Board 
of Governors to the Active grade: 



GANSELL, A. 

Audio Productions, Inc. 
250 W. 57th St. 
New York, N. Y. 

LOYE, D. P. 

Electrical Research Products, Inc. 
7046 Hollywood Blvd. 
Hollywood, Calif. 



ROESSLE, H. 

230 West End Ave. 
New York, N. Y. 



WOOD, C. R., SR. 
30-38 29th St. 
Astoria, L. I. 



JOURNAL 

OF THE SOCIETY OF 

MOTION PICTURE ENGINEERS 

Volume XXVIII APRIL, 1937 Number 4 

CONTENTS 

Page 
A Review of the Quest for Constant Speed E. W. KELLOGG 337 

The Effect of Aberrations upon Image Quality 

W. B. RAYTON AND A. A. COOK 377 

The Analysis and Specification of Color K. S. GIBSON 388 

Direct Recording and Reproducing Materials for Disk Record- 
ing A. C. KELLER 41 1 

New Motion Picture Apparatus: 
Three- Wire D-C. Supply for Projection Arcs. . .C. C. DASH 427 

Committees of the Society 431 

Spring, 1937, Convention; Hollywood, Calif., May 24th-28th, 
Inclusive 436 

Society Announcements 441 



JOURNAL 

OF THE SOCIETY OF 

MOTION PICTURE ENGINEERS 



SYLVAN HARRIS, EDITOR 

Board of Editors 

J. I. CRABTREE, Chairman 

A. N. GOLDSMITH L. A. JONES H. G. KNOX 

A. C. HARDY E. W. KELLOGG T. E. SHEA 



Subscription to non-members, $8.00 per annum ; to members, $5.00 per annum, 
included in their annual membership dues; single copies, $1.00. A discount 
on subscriptions or single copies of 15 per cent is allowed to accredited agencies. 
Order from the Society of Motion Picture Engineers, Inc., 20th and Northampton 
Sts., Easton, Pa., or Hotel Pennsylvania, New York, N. Y. 
Published monthly at Easton, Pa., by the Society of Motion Picture Engineers. 

Publication Office, 20th & Northampton Sts., Easton, Pa. 
General and Editorial Office, Hotel Pennsylvania, New York, N. Y. 

West-Coast Office, Suite 226, Equitable Bldg., Hollywood, Calif. 
Entered as second class matter January 15, 1930, at the Post Office at Easton, 
Pa., under the Act of March 3, 1879. Copyrighted, 1937, by the Society of 
Motion Picture Engineers, Inc. 

Papers appearing in this Journal may be reprinted, abstracted, or abridged 
provided credit is given to the Journal of the Society of Motion Picture Engineers 
and to the author, or authors, of the papers in question. Exact reference as to 
the volume, number, and page of the Journal must be given. The Society is 
not responsible for statements made by authors. 



OFFICERS OF THE SOCIETY 

President: S. K. WOLF, 250 W. 57th St., New York, N. Y. 
Past-President: H. G. TASKER, Universal City, Calif. 

Executive Vice-President, G. F. RACKETT, 823 N. Seward St., Hollywood, Calif. 
Engineering Vice-President: L. A. JONES, Kodak Park, Rochester, N. Y. 
Editorial Vice-President: J. I. CRABTREE, Kodak Park, Rochester, N. Y. 
Financial Vice-President: O. M. GLUNT, 463 West St., New York, N. Y. 
Convention Vice-President: W. C. KUNZMANN, Box 6087, Cleveland, Ohio. 
Secretary: J. FRANK, JR., 90 Gold St., New York, N. Y. 
Treasurer: L. W. DAVEE, 250 W. 57th St., New York, N. Y. 

GOVERNORS 

M. C. BATSEL, Front and Market Sts., Camden, N. J. 

A. S. DICKINSON, 28 W. 44th St., New York, N. Y. 

G. FRIEDL, JR., 250 W. 57th St., New York, N. Y. 

A. N. GOLDSMITH, 444 Madison Ave., New York, N. Y. 

H. GRIFFIN, 90 Gold St., New York, N. Y. 

A. C. HARDY, Massachusetts Institute of Technology, Cambridge, Mass. 

K. F. MORGAN, 7046 Hollywood Blvd., Los Angeles, Calif. 

C. H. STONE, 205 W. Wacker Drive, Chicago, 111. 



See p. 431 for Technical Committees 



A REVIEW OF THE QUEST FOR CONSTANT SPEED* 
E. W. KELLOGG** 



Summary. The importance of constant record speed in machines used for repro- 
duction of music was realized by Thomas A. Edison and many other pioneers in 
sound recording. Crude performance from other standpoints made it hardly worth 
while for the earlier workers to attempt to obtain extremely high standards of speed 
constancy. 

The flyball type of phonograph governor came into the picture and has been worked 
so well that it has not even yet been superseded, although with synchronous motor 
drives for certain types of equipment the governor is no longer necessary. 

Recording sound photographically probably began with the work of Alexander 
Graham Bell who made records on glass disks; but not until long celluloid films were 
available and the motion picture thoroughly established did photographic sound re- 
cording become a competitor with the disk. As late as 1930, there were many engi- 
neers who advocated the disk for sound-picture work. While the same general prin- 
ciple applies to both mechanical and photographic records, the latter involves certain 
additional problems. 

Among the earlier workers in this field, the expedients adopted by C. A. Hoxie 
and C. L. Heisler, of General Electric Company, deserve recognition. The paper 
gives brief descriptions and discussions of a number of ingenious arrangements for 
improving speed constancy that have been employed by various inventors and engi- 
neers. Some of these expedients have been applied to record turntables and some to 
film equipment. 

The need of constant speed in any device for recording or repro- 
ducing sound was quickly realized in the earliest experiments along 
these lines. Edison in patent No. 227,679, applied for in 1880, says, 
"In experimenting with the phonograph I discovered that the re- 
production of sound was imperfect if the slightest variation occurred 
in speed; hence the combination of a cylinder with a very heavy 
flywheel." 

Alexander Graham Bell and Sumner Tainter 1 experimented with 
various forms of phonographs. One of their early models is on ex- 
hibition in the Smithsonian Museum at Washington. The writer 

* Presented at the Fall, 1936, Meeting at Rochester, N. Y. ; received January 
29, 1937. 

** RCA Manufacturing Co., Camden, N. J. 

* 337 



338 



E. W. KELLOGG 



[J. S. M. p. E. 







ATTORNEY 

FIG. 1. Several types of centrifugal governors. 



April, 1937] THE QUEST FOR CONSTANT SPEED 339 

had occasion to have a copy made of this machine, which employed 
a large flywheel, and found that although this and the Edison machine 
were both driven by hand cranks, it was not difficult with the big 
flywheels to maintain a speed so nearly constant that fairly creditable 
reproduction took place. 

The fly-ball governor, constructed so as to introduce friction as 
soon as the speed exceeds a certain value, was employed in some of 
the earliest phonographs. Fig. 1 shows several forms of phonograph 
governor. This device worked so well that in one form or another it 
has remained an essential feature of practically all phonographs un- 
til the advent within the past five or six years of phonographs driven 
by synchronous motors. In an art such as sound recording and re- 
production, in which so many factors must be correct to produce a 
satisfactory overall result, it is not usual to find great refinements 
introduced in the effort to render one factor perfect so long as there 
are serious faults of other kinds. For this reason driving systems 
that today we might regard as short of satisfactory were in general 
considered good enough, until some of the recent improvements made 
the shortcomings in the matter of speed more evident. Examples of 
such advances are electric recording on wax, 2 ' 3 the "orthophonic" 
phonograph, 2 electric reproduction from disk, 4 improvements in am- 
plifiers and loud speakers, 5 - 6 ' 7 the development of record materials 
having lower surface noise, 8 the working out of satisfactory record- 
ing on film, and the extension of the frequency range 9 through recent 
advances in technic. Nevertheless, there were evidently engineers 
and inventors who, considerably before these improvements came 
into being, were conscious of forthcomings in the driving systems, 
and introduced the idea of filtering out irregularities in speed. A 
French patent No. 10,377, issued to the Fabrica Italiana, Pellicole 
Parlate, dated June 19, 1909, shows a record turntable driven through 
a spring, the spring being wound with tape in order to damp out 
oscillations. 

John Constable in a 1918 patent 10 showed a cylindrical record driven 
through springs, and gives an excellent description of the function 
of the springs in taking up the vibrations imparted by the driving 
system. It is possible that these inventors were stimulated to employ 
filtering systems because they did not have as good governors and 
gears as could be produced, for their devices did not come into general 
use until a much later date, and then in greatly modified form. Con- 
stant speed in recording or reproduction means not only uniformity 



340 



E. W. KELLOGG 



fj. S. M. p. E. 



of rotation of the record, but avoidance of other relative movements 
of the record and pick-up device. Thus, the effect of vibration of 
the pick-up, or record, or both, is the same as if there were changes in 
rotational speed of corresponding frequency and amplitude. E. H. 
Amet in a 1917 patent 11 relating to sound motion pictures, shows a 
very well thought out system for preventing the transmission of 
vibrations from his picture projector to his phonograph. 

APPLICATION OF FILTERS TO TURNTABLES 

With the need of lower-speed turntables for sound motion pictures, 
the difficulties of securing constant speed were greatly multiplied. 




R-l 



FILTER WITH DAMPED SPRING 



ij REPRESENTS DISTURBANCES AT TURNTABLE.. 

Ii REPRESENTS DISTURBANCES INTRODUCES BY GEARS OR DRIVING SYSTtM- 

C Z REPRESENTS DISTURBANCE* INTRODUCED BK BEARING FRICTION * NEEDLE PRA6. 




RELATIVE FREQUENCY 

FIG. 2. Characteristics of commercial type of filter. 

The difficulties were further increased by the necessity of synchronous 
operation, which meant geared drive. 

Although it is possible to produce gears having a very high order of 
precision, it is practically impossible to drive any mechanism through 
gears without the introduction of some vibration of the tooth fre- 
quency. As soon as attempts are made to build geared machines in 
any reasonable numbers and with reasonable manufacturing toler- 
ances, it has almost invariably turned out that complete dependence 
could not be placed upon the perfect functioning of gears. The 
errors may be small but they are sufficient to impair sound quality. 
It is comparatively easy to eliminate the tooth vibrations by driving 
the turntable, which has considerable moment of inertia, through a 



April, 1937] THE QUEST FOR CONSTANT SPEED 341 

flexible element. Such an arrangement is called a mechanical "filter." 
Fig. 2 shows the characteristics of a simple filter comprising a single 
mass and a single elastic element, for example, a turntable driven 
through a spring. The ordinates show the ratio of amplitude of 
swing or speed variation of the turntable to that of the driving gear 
at the other end of the spring. For disturbances of a frequency at 
which this ratio is greater than unity, the filter does more harm than 
good. It is evident from Fig. 2, that damping must be provided, 
since otherwise a bad resonance occurs and persistent oscillations are 
encountered. Damping may be applied to the spring, or to the turn- 
table. Both systems have been employed, but for the most part the 
damping has been applied to the spring. 




FIG. 3. Laminated worm-gear of Elmer and Blattner. 

It will be noted that the filter is of no use except when the fre- 
quency of the disturbance is above a certain rate. The successful 
employment of a filter therefore depends upon avoiding, so far as 
possible, disturbances of slow periodicity, and then designing the 
filter so that it will be effective at the lowest frequency that must 
be expected; which means making the spring flexible enough 
and the turntable heavy enough. This is often spoken of as giving 
the filter a "low cut-off frequency." Too often the design has been 
guessed rather than calculated. H. C. Harrison in U. S. Patent 
No. 1,847,181 gives a very complete formulation of the requirements. 

The lowest disturbing frequency is generally the rotation rate of 
the slowest gear of the train, and with motor-driven machines this 






342 



E. W. KELLOGG 



[J. S. M. p. E. 



can be the turntable speed. When the turntable speed itself is as 
low as 33 Va r P m - ( tms l w s P ee d having been adopted to provide 
long playing records for sound picture work) filtering presents serious 
problems. An interesting arrangement for reducing the magnitude 
of the disturbances to be filtered out, and at the same time increasing 
the frequency of their recurrence, was described by Elmer and Blat- 
tner in 1929. 12 It is assumed in considering the principle of this 
driving system, that practically all the disturbance of fundamental or 
turntable frequency will be due to imperfections in indexing the 
gear on the turntable spindle. Such indexing imperfections are 
primarily traceable to the master gear in the miller upon which the 
gear is cut, and there is no way of averaging out these imperfections 
in the cutting operation. The imperfections in the gear as cut, 




FIG. 4. Diagrams illustrating averaging linkage described by 
Elmer and Blattner, and straight-line analogue. In actual struc- 
ture, each gear lamina carried two pins, and equalizing links were 
in duplicate. 

however, can be averaged by the scheme described by Elmer and 
Blattner. The gear consists of four laminae which are bolted to- 
gether and milled and hobbed in the usual manner. The four laminae 
are then separated and each rotated 90 degrees with reference to the 
adjacent ones. They are mounted in such a way that slight play 
between layers can occur without excessive friction. This compound 
gear is driven from a single worm. Fig. 3 shows the compound gear 
with worm. It will be appreciated that if a certain imperfection in 
the indexing tends to cause one layer to rotate above average speed, 
the same effect will not be produced upon the layer that is reversed 
with respect to the first layer until the gear as a whole has rotated 
180 degrees. Thus, although the speed of the individual layers is no 
more nearly constant than that of a single gear, the average speed of 
the four layers will be much more nearly constant, and such irregu- 



April, 1937] 



THE QUEST FOR CONSTANT SPEED 



343 



larities as occur in the average speed will repeat themselves four 
times per revolution. 

The system for averaging the speeds of the four laminae is analo- 
gous to the action of a whiffletree which provides an equalization of 
the motions of several horses, as illustrated in Fig. 4. On each 
lamina are mounted two vertical pins which project through slots in 




FIG. 5. Planetary turntable drive of A. V. Bedford. 

the laminae above, thus providing eight points of attachment for the 
linkages which equalize the movement. Fig. 4 shows the equalizing 
linkages in schematic form. The turntable is driven through springs, 
while an oil dash-pot having radial vanes provides the damping. The 
dash-pot is driven from the linkage. Thus, filtering is employed in 
addition to the speed-averaging device just described, which reduces 
the sources of disturbance that must be filtered out. 

Bedford Planetary Drive. A radically different method of increasing 
the frequency of repetition of such disturbances as may be due to 



344 E. W. KELLOGG [J. S. M. P. E. 

faulty gears has been employed by A. V. Bedford 13 and is illustrated 
in Fig. 5. On the turntable spindle is a gear E, and below this is a 
stationary gear /, of slightly smaller diameter. Two pinions F and 
U, which are coupled together and rotated by a planetary arrange- 
ment, engage gears E and /. Were these gears of the same diameter, 
no rotation of E would result; but since E differs in diameter from 
/, each revolution of the pinion-carriage G results in a slight rotation 
of the gear E. The diameters are so chosen that the pinions must 
make about seven revolutions to produce one revolution of the turn- 
table spindle. In the course of this one revolution, all parts of gear E 
have been engaged six times and all parts of gear / seven times. In 
order further to reduce the possibilities of imparting low-frequency 
disturbances to the turntable, its spindle rotates not in a stationary 
bearing but in a bearing that rotates with the planetary pinion 
carriage G. 

Any torque imparted to the gear E reacts upon the gear /. Were 
the latter not restrained it would rotate instead of gear E. Instead 
of providing a rigid anchorage for the stationary gear 7, it is restrained 
by means of a spring L, and damping is provided by the dash-pot K. 
If the turntable is designed with a reasonable moment of inertia, 
irregularities in the speed of gear E with reference to gear /, such as 
may result from gear imperfections, tend to produce movements of 
gear I rather than speed fluctuations of the turntable gear E. Since 
the moment of inertia associated with gear / is quite small, / takes 
up the irregularities, thus providing the filtering upon which reliance 
is placed to eliminate the comparatively high-frequency disturbances 
that the planetary drive may introduce. 

Brute- Force Methods. It may seem paradoxical that such elaborate 
arrangements and complicated constructions as are here described 
should be necessary in order to provide constant motion when, ac- 
cording to the laws formulated by Isaac Newton, all that is necessary 
to cause an object to move with uniform velocity is to let it alone. 
Fortunately for disk record work, the power requirements are ex- 
tremely low, and only such considerations as quick starting, saving 
of space or weight, or the necessity of exact synchronism, make it 
necessary to resort to any other expedients than the closest prac- 
ticable approach to "letting it alone." 

The employment of very large moment of inertia is often described 
as a "brute-force" method, but this approach to the problem is good 
only when extreme care is exercised at the same time to apply a 



April, 1937] THE QUEST FOR CONSTANT SPEED 345 

minimum of forces to the rotating mass. Fig. 6 shows the turntable 
that comes the nearest to constant speed of any with which the au- 
thor has had acquaintance. The flywheel is 27 inches in diameter, 
with a heavy rim. The spindle runs in a carefully made sleeve 
bearing, and the weight is carried by a steel ball at the bottom. 
Careful bearing design minimizes the disturbing forces from this 
source, and the supply of the power through a long thread minimizes 
disturbing forces from the driving system. 

Measurement of Speed Fluctuations. The testing of turntables for 
speed constancy has involved even more serious difficulties than the 
provision of a turntable having a minimum of speed fluctuations. 




FIG. 6. Turntable used for reference standard. 

Stroboscopic methods become increasingly unsatisfactory as the 
frequency of the disturbance increases. A better system consists 
in causing the apparatus under test to generate an electric current, 
the frequency of which depends upon the speed of running. This 
alternating current is applied to a circuit that is slightly off tune, and 
the variations in frequency then appear as variations in voltage which 
can be measured and recorded. 14 For generating the tone it has not 
been found altogether satisfactory to cut a record and play it back. 
Better results have been obtained by means of a carefully constructed 
magnetic tone-wheel. The problems of measurement of speed 
fluctuations in turntables is discussed by A. R. Morgan and the 
writer in a paper published in the journal of the Acoustical Society 
of America for April, 1936. Using the magnetic tone- wheel on the 
turntable shown in Fig. 6, we have found it possible to get registra- 



346 



E. W. KELLOGG 



[J. S. M. p. E. 



tions of as low as 0.1 per cent between the lowest and the highest 
speeds registered in a period of 6 seconds. The measuring system 
shows full sensitivity to fluctuations up to 100 per second. 



SOUND-FILM SYSTEMS 



Photographic recording of sound on film began a good many years 
ago. The problem of obtaining constant speed does not differ in 




FIG. 7. Example of "brute-force" constant-speed film 
phonograph. 

fundamental principles from that of providing constant speed for 
disk records, but does present new problems in that the entire record 
does not move as a unit but is unwound from a supply reel, pulled 
past the point where a record is made or reproduced, and again wound 
upon a reel. The motion at the translation point must be protected 
from irregular pulls due to the reels. An additional difficulty is 



April, 1937] 



THE QUEST FOR CONSTANT SPEED 



347 



generally brought in by the requirement that a specified number of 
sprocket holes must pass per second, regardless of the absolute length 
of film represented by this number of sprocket holes. In other words, 
the linear speed for a shrunken film must be less than for a non- 
shrunken film. Apparent lack of concern for speed constancy upon 
the part of earlier experimenters in photographic sound recording is 
probably due to the fact that they had too many worries about other 
things. That the speed should be constant, however, has always 
been accepted as an axiom. Machines employing the "brute-force" 
method of moving a film at constant speed have been of extreme value 
as laboratory tools. Fig. 7 shows such a machine, known in the RCA 
Photophone laboratories as the ^^^^ 

"Grindstone." Such extremes of 
size and mass are hardly neces- 
sary. In general, rotational 
speeds in film machines can be 
considerably higher than in disk 
turntables (for example, 180 to 
360 rpm.), and at these speeds 
flywheels of moderate size are 
sufficient. 

Filters on Sprocket Shaft. In 
most of the earlier commercial 
reproducing machines, and in 
many of the recording machines, 
the film was moved past the 
optical system by means of a sprocket; and, practically without 
exception, the sprocket drive has been filtered. In other words, a 
flywheel was mounted upon the sprocket shaft, and power was 
supplied through an elastic connection from a driving gear. As has 
already been explained, a filter may do more harm than good if 
it is not adequately damped. The damping has in some cases 
been applied to the flywheel itself in the form of a viscous brake, 
and in other designs the elastic connection between the drivirg 
gear and the flywheel was damped. Fig. 8 shows a widely used 
damping arrangement employed in projector sound heads. 16 The 
two sylphons, or bellows, which are mounted upon the flywheel are 
completely filled with oil and connected through an adjustable orifice. 
The mechanism is so arranged that whenever there is relative move- 
ment of the flywheel and driving gear, one bellows is stretched and 




FIG. 8. Sylphon damping arrangement 
of H. Pfannenstiehl. 



348 



E. W. KELLOGG 



[J. S. M. p. E. 



the other is compressed, and oil must flow between them. A device 
of this form provides very effective damping but must be designed 
with great care in order that changes in the torque transmitted may 
not result in appreciable unbalance. All filtered systems with hori- 
zontal axes are extremely sensitive to balance. 

Inherent Faults in Sprocket Drive. Although the painful effects of 
slow speed fluctuations, particularly in the reproduction of music, 
were early observed, and efforts made to eradicate them, general 
recognition of the fact that rapid fluctuations of small amplitude can 
also do serious damage to quality has been extremely tardy. This 
is perhaps due to the fact that such rapid fluctuations produce radi- 

(CIRCULHR 
\-INrOLUTl 




FIG. 9. Common relation between sprocket teeth and perforations. 

cally different effects, which were not recognized as being due to 
speed changes. Moreover, other common faults in recording and 
reproduction cause impairment of quality so similar to that due 
to rapid speed variations as to mask the results in comparative tests. 
The art of sound reproduction may be metaphorically described as 
"strewn with the wrecks of efforts" to obtain constant speed by means 
of gears and by means of sprockets. The shortcomings of sprocket 
drives are made much more serious because the pitch of the film per- 
forations varies with the shrinkage of the film. No matter how the 
sprocket is designed, there is only one chance in many that the film 
will really fit the sprocket. 24 

Fig. 9 shows the relation between teeth and sprocket holes when 
the pitch of the perforation is slightly greater than that of the teeth. 
It is seen that the tooth on the left is doing the pulling and will con- 



1937] 



THE QUEST FOR CONSTANT SPEED 



349 



me to do so until the film is stripped off, at which time the film must 
slip back on the sprocket by a distance S; whereupon tooth No. 2 
does the pulling. Thus, the film moves at sprocket-tooth speed only 
if or short intervals, interspersed with moments of slipping back. 
If the sprocket-hole pitch is less than the tooth pitch, there must be a 
forward adjustment as each new tooth engages, instead of a slipping 
iback. In either case, there is a decided irregularity in film move- 
ment, the effect of which is to cause less satisfactory reproduction of 
ihigh-frequency tones. 

Fig. 10 shows what happens to tones of several different frequencies 
las a result of speed fluctuations at sprocket-hole frequency. 16 - 17 If 
ia sine wave is recorded and is reproduced by a machine that intro- 



200 300 * 



FIG. 10. Effect of film speed modulation upon sine wave recordings of 
300, 1000, 2000, 4000, and 6000 cps. (Assumed: speed modulation 
of 1 mil amplitude at 100 cps.) 

duces a variation of 0.001 inch in amplitude approximately 100 
times per second, there is little effect upon a 300-cycle tone, although 
small amounts of 200- and 400-cycle tones are generated. If the 
recorded frequency is 1000 cps. the magnitude of the 900- and 1100- 
cycle tones is about 15 per cent of the original. At 2000 cycles the 
i first sideband tones are fully 30 per cent of the original, while addi- 
j tional tones of 1800 and 2200 become appreciable. If the recorded 
tone is 4000 cps., the sidebands are practically as large as the funda- 
mental, while at 6000 cps. there are four sidebands that exceed the 
fundamental and two more that are nearly equal to the fundamen- 
! tal. It could hardly be expected that a recorded 6000-cycle note re- 
' produced on such a machine would sound clean and clear, although 
some semblance of a high-frequency note still results. 

Since satisfactory reproduction can not be expected if the film is 



350 



E. W. KELLOGG 



[J. s. M. p. E. 



propelled past the optical system directly by a sprocket, the best 
machines have employed smooth drums on which the films were 
supported, dependence being placed upon friction to insure the 
film's moving with the drum without slipping. Soft-tired rollers 
pressing the film against the drum have proved satisfactory and make 
little trouble if properly designed. Various methods of driving the 
drum have been employed, and among these methods it is regrettable 
that there have been many attempts to drive the drum by gearing 
it to the same driving system as the sprocket, without provision for 
permitting the drum to operate at any other than a fixed speed with 



Approximate 

Adjustment (Manual) 




FIG. 11. Electrical drum speed control of C. A. Hoxie. 

reference to the sprocket. Such machines may have the appearance 
of operating as if the drum were performing a useful function, but 
as a matter of fact, the film must be slowly slipping with reference 
to the drum surface throughout the operation, and all benefit from 
the constant-speed drum is thereby sacrificed. The drum must be 
permitted to select its own speed, depending upon the film shrinkage. 
The simplest way to let the drum select its own speed is to drive it 
by means of the film, which acts as a belt. In applying this prin- 
ciple, dependence has almost always been placed upon a flywheel to 
provide for uniform rotation of the drum. It would be hard to 
imagine a simpler system that would eliminate the disturbance due 
to sprocket teeth. The sprocket tooth disturbance present at the 
sprocket pulling the film is filtered out because of a certain amount of 



April, 1937] 



THE QUEST FOR CONSTANT SPEED 



351 



flexibility in the loop of film between the sprocket and the drum 
cooperating with the effective mass at the drum due to the flywheel. 
The trouble with the simple drum and flywheel system is that it 
constitutes an undamped filter such as was discussed in connection 
with Fig. 2. Oscillations are easily set up which persist for a long 



June 2, 1931. 

1808551 



C. L. HEISLFR 
fllH DRIVING 

1927 




Drum 



Drum Shaft 
Flywheel 



Intermediate 

Friction TTheel 
Driven 

Friction Wheel 



FIG. 12. 



Mechanical drum speed control or compensator 
of C. L. Heisler. 



time. A small amount of damping is, to be sure, provided by the 
viscosity of the bearing lubricant; and if through careful construc- 
tion, or good design, or good luck (or, perhaps, by tedious and painful 
reconstruction), no disturbance of frequency near the natural fre- 
quency of the oscillatory system is introduced, a machine of this 
type may give creditable performance. A number of other methods 
of driving the drum at the correct speed have been employed, and 
some of these are of much interest in view of their ingenuity. 



352 E - W. KELLOGG [j. s. M. p. E. 

Fig. 11 shows the arrangement of an early laboratory reproducing 
machine of Charles A. Hoxie, of the General Electric Co. The drum 
is driven by a separate d-c. motor, the speed of which is controlled 
through a fine-adjustment rheostat, the position of which is changed 
by alterations in the lengths of the loops between the sprocket and 
the drum. 18 This system is not subject to oscillation, since the 
motor assumes its new speed almost instantly when the rheostat is 
changed. It is in systems in which the acceleration of the drum 
(rather than its speed) varies in proportion to film loop length, that 
oscillations are likely to occur. 

C. L. Heisler, also of the General Electric Co., designed a recorder 
that was widely used commercially, using a mechanical method of 
introducing the fine control of the drum, instead of an electrical 
method. 19 Fig. 12 shows some of the features of the Heisler ma- 
chine. The principle is in no wise different from the employment of 
two oppositely tapered cones, with a belt or friction wheel that can be 
moved axially to the position at which the speed ratio is correct. 
Instead, however, of simple conical pulleys, Mr. Heisler employed 
doubly curved surfaces of revolution, A and C of Fig. 12, the ele- 
ments of which were circular arcs. This made it possible to rotate 
the axis of the intermediate friction wheel B, instead of shifting the 
wheel along its axis. The intermediate friction wheel was mounted 
much like a gyroscope, except that its axis had only one degree of 
freedom. On the drum side, the running surface is on the interior 
of a hollow member C. It will be observed that as the axis of the 
intermediate wheel B is rotated in a clockwise direction, its rim is 
brought into contact with a larger diameter portion of the driving 
cone A; and, on the other side, it runs against a part of the driven 
member C where the diameter of the latter is less. This provides a 
continuous adjustment of speed. The upper drawing of Fig. 12 
shows the loop of film between the sprocket of the drum, engaging a 
movable idler at the bottom. Changes in the position of the axis of 
the friction wheel are caused by the lengthening and shortening of 
this loop of film. 

Geared Compensator of C. L. Heisler. Fig. 13 shows another arrange- 
ment by Heisler. 20 A ring-gear, A, is mounted on the shaft whose 
speed is to be controlled, and a slightly larger ring-gear, B, is driven 
at fixed speed from the motor. Two gears C and D, mounted on a 
single hub, which the writer has called "creeping gears," mesh with 
the two ring gears. The creeping gears run on an eccentric hub E. 



April, 1937J 



THE QUEST FOR CONSTANT SPEED 



353 



When this hub is permitted to rotate with the rest of the assembly, 
there is no relative motion of the four gears, and the driving and 
driven ring-gears rotate at the same speed. Application of a brake, 
however, to the eccentric hub causes the pair of creeping gears to run 
around inside the internal gears, and in doing so they cause one to 
shift slowly with respect to the other. 

The smaller the difference between the diameters of the two in- 
ternal gears, the smaller is the creeping produced by applying the 
brake to the eccentric hub, and the arrangement may be so designed 
that the entire difference in speed of the driven shaft produced by 
applying a brake is only 2 or 3 per cent of the mean speed. 

The drawing on the left in Fig. 13 shows the manner in which the 




CKCCPIHG GfAKS 
SCHEMATIC REPRESENTATION OF GEARED COMPENSATOR OF CL HEELER 

FIG. 13. Geared compensator of C. L. Heisler. 

loops of film between the sprocket and the drum control the applica- 
tion of the brake. This drawing shows the arrangement as if the 
geared compensator were applied to alter the speed of the drum, but 
in most of the machines that Heisler designed it was not the speed of 
the drum that was changed but the speed of the rest of the machine, 
comprising all the sprockets. This is entirely permissible in a pro- 
jector, which does not have to be synchronized with other machines. 
Since the geared compensator is a means of controlling average speed, 
but does so by causing small fluctuations, it is logical to drive the 
drum directly through gears and a filter from the motor, and to employ 
the compensator to insure that the sprockets will have the correct 
average speed to maintain proper film loops. In such a machine it 
is evident that the film speed in linear feet per minute is the fixed 
quantity, while the number of sprocket holes per minute will vary, 
depending upon film shrinkage. 



354 



E. W. KELLOGG 



[J. S. M. p. E. 






In an alternative construction, the creeping of one gear with 
respect to another was brought about by an arrangement of worms 
that rotated with the assembly but did not turn upon their axes un- 
less a brake was applied to a suitably arranged brake-wheel. In 
practice Heisler found that the action of the brake was such as to 
cause the brake-wheel to assume an intermediate speed rather than 
to alternate between full speed and stop. This was due to the facts 
that the brake-wheel possessed considerable inertia and the entire 
mechanism operated under conditions of abundant lubrication so 



Reduced Cross Section 
of Rubber and Consequent 
Higher Velocity 




Region of High' Velocity 



Water Pipe Analogue 

FIG. 14. Speed control by pressure on rubber tire, used 
by C. A. Hoxie. 

that the brake produced a semi- viscous drag. Since the total speed 
change produced by stopping the brake-wheel amounted to only 
2 or 3 per cent of the average, it is evident that even a rough ap- 
proximation to constant brake-wheel slip would result in a very nearly 
constant speed of the driven member. Moreover, filtering may be 
employed between the driven member of the compensator and a 
drum. This method of operation permits the application of the 
geared compensator to machines such as recorders, which have to be 
strictly synchronized. 

Rubber-Tired Compensator of C.A. Hoxie. Hoxie built a sound head 
in which he provided continuous fine adjustment of the drum speed 



1 April, 1937] 



THE QUEST FOR CONSTANT SPEED 



355 



by altering the pressure between a rubber-tired roller, driven at con- 
stant speed, and a smooth-rimmed metal wheel. 21 I am not sure 
how Mr. Hoxie expected the device to work when he made the first 
model, but it may be said that a number of us expected exactly 
the opposite from the actual result. The working radius of the 
driving wheel becomes less as the tire is compressed. We therefore 



April 8, 1930. 



A V BEDFORD 

BELT OR STRIP DRIVING ARRANGEMENT 
Filed Nov 3. 1927 



1,754,187 




Fig 2. 




FIG. 15. Film speed control of A. V. Bedford. 

inferred that the driven wheel would run slower as the pressure was 
increased, assuming, of course, that the driven wheel rotated easily 
and that the pressure was never reduced to a point where appreciable 
slipping occurred. What actually happened was that the greater 
the pressure, the faster the driven wheel ran. 

The proposition may be illustrated in terms of a truck. We shall 
assume that the pavement is smooth and level, and that the truck 
has thick, soft tires of solid rubber, and that the diameter of the 
rear wheels is such that when they are rotating at 200 rpm. the truck 



356 E. W. KELLOGG [J. S. M. P. E. 

will run 20 miles per hour when empty. The truck is now loaded 
sufficiently to compress the tires materially. With the same wheel 
rpm., will the loaded truck run slower or faster than it ran when 
empty? The answer is that the more you load it the faster it will 
go. I know of no actual test with a truck, but believe that the 
analogy is a fair one, and the explanation is quite simple. 

Fig. 14 illustrates the relation of the rollers and also shows a dia- 
gram of a Venturi tube. For such changes of shape as a piece of 
soft rubber can undergo it acts much like a liquid, in that it permits 
large shearing deformations with little resistance. It is evident that 
at the point of maximum compression, the tire has smaller cross- 
section than at other points, but it is also obvious that the total 
volume of rubber passing this point in a given interval is the same as 
that passing other points. Therefore the velocity of the rubber 
must be greater where the tire is compressed, and the greater the 
compression, the higher will be the velocity of the rubber, just as the 
velocity of water at the constricted portion of the Venturi tube is 
greater than that in other parts of the pipe. Directly under the 
truck wheels the surface velocity of the rubber with reference to the 
truck may considerably exceed the velocity of other parts of the tire 
surface, and this effect much more than offsets the slight reduction 
in radius of action. If the tire were compressed to J /2 its normal 
cross-section, the velocity of the rubber at this point would need to 
be doubled, on the average; and since this increase in average ve- 
locity will be mostly at the surface and none at the rim, very sub- 
stantial increases in speed may occu*. 

Pinch-Roller Compensator of A . V. Bedford. A method of compen- 
sation that has surpassed even the rubber-tire compensator in its 
ability to provide subject matter for discussion and argument, and 
likewise resembles the rubber-tire arrangement in its extreme sim- 
plicity, is the pinch-roller compensator, first proposed by Bedford 
for projectors 22 and later applied to printers. 23 

In Fig. 15, the upper drawing shows the general mechanical ar- 
rangement, while the lower drawing shows the principal elements in 
simplified form. Roller 4 is driven at fixed speed, while roller 5, 
which serves to hold the film in contact with roller 4, may assume 
whatever speed the film imparts to it. If we could stretch or com- 
press each film to such a degree that 64 sprocket holes corresponded 
to exactly one foot, thereby mechanically compensating for shrinkage, 
it would obviously be possible even in a synchronous machine to 



April, 1937] THE QUEST FOR CONSTANT SPEED 357 

propel film by means of a smooth roller having a fixed peripheral 
speed. Fundamentally it is not necessary to stretch or compress 
the film throughout its entire cross-section, for the film is driven 
solely by contact between one surface and roller 4. It is therefore 
sufficient to stretch or compress the material at this surface, and it 
is well known that the simple act of bending a flexible strip serves 
to compress one surface and stretch the other. It will be readily seen 
from the drawing that if a sufficient amount of film accumulates in the 
loop between the sprocket and the driving roller, the film will bend 
around the roller in such a way that the latter works on the concave 
side of the bent film. On the other hand, with a short loop, as indi- 
cated at 19, the driving roller will operate on the convex side of the 
bent film. When the driving is applied to the convex side, the mean 
film speed is less than that of the driven surface. In other words, 
when the loop is short the film does not pass so rapidly through the 
driving point between rollers 4 and 5. Whatever the degree of 
shrinkage of the film, an equilibrium is soon reached with the loop 
length such that the rate of passage between the rollers is exactly 
equal to the rate at which it is fed through by the sprocket. Engi- 
neers dealing with problems of belt drive are well acquainted with 
the necessity of making a distinction between pulley surface speed 
and mean belt speed, or the speed of the middle of the belt, and the 
Bedford drive depends upon exactly the same principle. 

THE MAGNETIC DRIVE 

The several systems just described, in which the drum is positively 
driven and compensation provided, possess one advantage over the 
system in which the drum is driven by the film only; namely, more 
rapid acceleration may be permitted, since the strength of the film is 
not a limiting factor. In most equipment, however, extremely 
rapid acceleration has not been essential, and in the case of the mag- 
netic drive used in RCA recorders, the film is required to perform only 
a small part of the work of accelerating the drum flywheel. 

During the period when the numerous fundamental problems of 
sound-on-film were being worked out, there were, of course, many 
machines both for recording and reproducing that fell seriously short 
of an adequate standard of constant speed. The recorder rather 
than the reproducer appeared to be the logical place to apply serious 
efforts to build a machine that would be less subject to speed fluctua- 
tions. In the machines, of which a number were built, that depended 






358 



E. W. KELLOGG 



[J. S. M. P. E. 



upon the film to pull the drum, there were two causes of failure to 
operate satisfactorily. In the first place, the tension on the film re- 
quired to keep the drum running was sufficient to pull the film quite 
tight between the pulling sprocket and the drum, and this taut loop 
provided too little flexibility. This made it impossible to design the 
filtering system so that it would have the necessary low cut-off fre- 
quency, the importance of which has been discussed in an earlier 
part of this paper. A heavier flywheel would tend to lower the filter 
cut-off frequency, but all possible gain from heavier flywheels was in 
turn lost by the increased film tension needed to drive the heavier 




FIG. 16. Film path of original magnetic drive recorder. 

system. In the second place, there was no provision for damping out 
oscillations. L. T. Robinson had introduced into such a machine a 
certain amount of damping by mounting upon the shaft a copper 
disk which rotated between the poles of a permanent magnet. 24 The 
result may have been helpful, but did not go far enough in damping, 
while it caused a further increase in the power required to rotate the 
drum. With this picture of the general problem, the writer decided 
to attempt to build a recorder in which these sources of trouble would 
be avoided. Attempt was made at the same time to avoid all other 
foreseeable sources of speed fluctuation that seemed likely to occur. 
In the machine as built there were some provisions that turned out to 
be unnecessary, but if one wishes to make sure that his horse will not 
be stolen, he may go so far as to lock the barn door, board up the 



April, 1937] THE QUEST FOR CONSTANT SPEED 359 

windows, place guards around the barn, and in addition to the fore- 
going, he may remove his horse from the barn altogether and keep 
him in his cellar until the scare is over. Fig. 16 shows schematically 
the arrangement employed. 

The sprocket that controls the speed at the drum is protected from 
jerks produced by the magazine take-up by a second, or buffer- 
sprocket geared to the driving system. Although it was planned 
to make the damping magnet help drive the flywheel rather than 
impose an additional drag, and thereby to make a more flexible film 
loop possible, there was no ready way of predicting whether the 
desired amount of film loop flexibility could be obtained by this ex- 
pedient alone. A movable idler roller as shown at M in Fig. 16 was 
therefore introduced in the loop between the pulling sprocket and 
the drum. By this means a large amount of flexibility can be intro- 
duced. An oil dash-pot connected to the movable idler provided as 
much damping as was required to make this part of the system 
practically dead-beat. 

Although the introduction of the movable idler provided flexi- 
bility and damping, there was some doubt as to whether it would 
provide the necessary filtering for rapid pulsations in the movement 
of the film at the sprocket. The effectiveness of the roller for this 
purpose would be impaired by its mass (although it was made of 
duralumin, and as light as practicable) and because the attachment 
of the dash-pot would increase the difficulty of its executing quick 
movements. It was therefore desirable to provide as much addi- 
tional flexibility as could be obtained in the film loop itself. It 
appeared to the writer that if the film were passed around two 
adjacent rollers in the same direction, it would assume a more 
rounded curve between them, and would therefore show more flexi- 
bility at the same tension than if the more common arrangement 
were followed of placing consecutive rollers upon opposite sides of 
the film, giving 5-shaped bends. Subsequent measurements indicated 
that this was true in even greater degrees than had been anticipated. 

Two soft-tired pressure-rollers were employed to hold the film 
tightly against the drum. It is, of course, especially important that 
the film be wrapped tightly around the drum where it passes the 
optical system. Tension on the film is obviously conducive to the 
desired tight wrapping. A certain amount of drag on the upper 
pressure-roller adds to the tension on the film where it passes around 
the drum, but drag by the lower pressure-roller tends to subtract 



360 



E. W. KELLOGG 



fj. S. M. P. E. 



from the tension, or would do so if any slipping on the drum occurred. 
To secure the best possible conditions in this respect, the upper and 
lower pressure-rollers were geared together through a slipping clutch in 
such a manner that power was taken from the upper pressure-roller and 

Dec. 27, 1932. E. w KELLOGG 1,892,554 

FILM SUPPORTING AND DRIVING APPARATUS 
Filed July 27, 1928 




FIG. 17. Schematic illustration of magnetic drive. 

delivered to the lower pressure-roller, thus providing a braking action 
above and at the same time applying a forward torque to the lower 
pressure-roller. The gear ratio was such that only slight slip occurred 
in the clutch. Most of the power absorbed by the upper roller was 
therefore delivered back to the drum by the lower roller, the purpose 
of this being to minimize the power that would have to be supplied to 
the drum. 



April, 1937] 



THE QUEST FOR CONSTANT SPEED 



361 



In one respect the original design deserves criticism. We have 
since that time learned more about edge guiding, but the faults of 
this machine in that respect were soon overcome and its value was 
all the greater because it served to demonstrate one or two things 
not to do. The guiding problem is discussed in some detail in an 
earlier paper 24 describing the magnetic drive recorder. 

It is well known that the reaction of a magnet upon a conducting 
disk of non-magnetic material is proportional to the relative velocity. 




FIG. 18. Electrical analogue of magnetic drive. 



Analogies: 

Ii, film movement at sprocket. 

/z, film movement at drum. 

LI, mass of flywheel. 

2, mass of magnet. 

Cj, flexibility of film loop. 

Ei, steady pull of film. 

2, power supply to magnet. 

Di, disturbances in sprocket drive. 

D t , variations in film stiffness, splices, 
kinks, etc. 

DI, irregularities in drum shaft bear- 
ing friction. 

Z?4, irregularities inmagnet drive gears. 



RI, coupling between magnet and 
flange. 

J?2. bearing oil viscosity. 

Ci, flexibility of spring attached to 
movable idler. 

Lj, mass of roller and arm. 

R it dash-pot resistance. 

Cz, spring between roller and dash- 
pot, if spring is used (33, Fig. 17). 

S, connection substituted for magnet 
drive, to make diagram applicable 
to rotary stabilizer. 



The damping properties of such an arrangement are proportional to 
the steepness of the curve of torque vs. relative speed. The damping 
is therefore independent of the continuous torque resulting from any 
constant difference in speed. The purpose of the design was to pro- 
vide enough damping to prevent oscillations, and only so much steady 
torque as would be needed to overcome bearing friction, and thus 
relieve the film of most of its load. The magnets were driven faster 
than the drum so that the continuous torque would be in the direc- 
tion to help rotate the drum rather than retard it. About 15 per 



362 



E. W. KELLOGG 



[j. S. M. P. E. 



cent excess magnet speed was selected. This figure was a guess, but 
proved to be about right. In subsequent designs it has been found 
feasible to reduce the excess magnet speed to about 8 per cent, which 
change permits increasing the damping. 

The ideal magnet arrangement for producing an eddy current 
drag on a disk or flange would consist of closely spaced and oppositely 
poled magnets, producing a flux that zigzags through the flange. 
Such a system, however, is much more difficult to arrange than one 
in which the flux through the flange is all in the same direction but 




FIG. 19. Front view of original magnetic drive recorder. 

concentrated at certain points by the shape of the poles. The latter 
arrangement was therefore chosen, and again the design, which was 
very much of a guess, seems to have been fairly well vindicated. 

Consideration was given to the production of the necessary rotating 
magnetic field by means of a polyphase winding on a stationary lami- 
nated field structure. The choice in this instance can be based 
upon simplicity of construction or operating convenience. The 
employment of a polyphase winding would require a power supply of 
the correct frequency, whereas when the excitation is by direct cur- 
rent, a battery supply is always obtainable, and the correct rota- 
tional speed is provided by whatever source of power operates the 
driving motor. 



April, 1937] 



THE QUEST FOR CONSTANT SPEED 



363 



Fig. 17 shows schematically the arrangement of the magnet drive. 
The figure also shows a refinement in the construction of the dash- 
pot and its associated idler. These are coupled together through a 
small spring 33, which permits the idler to execute quick movements 
of small amplitude without encountering the restraining action of 
the dash-pot. On the other hand, if the stiffness of the springs 33 
and 30 are correctly proportioned, the dash-pot is still effective for 
damping low-frequency oscillations, and it is only at moderately low 




FIG. 20. Rear view of magnetic drive recorder, showing magnet. 



frequency that the system has any tendency to oscillate. The elec- 
trical analogy of the drive and filtering systems is shown in Fig. 18. 
This diagram follows the usual conventions. A source of distur- 
bance is indicated by the symbol for an alternator, sources of con- 
tinuous torque by batteries, and speed corresponds to current. Bear- 
ing friction is represented by a resistance, for the component due to 
oil film viscosity, a counter emf. cell, for rubbing friction, and an 
alternator to imitate irregularity in the latter. It will be seen that 
by giving E% the necessary value, the necessary direct current can be 
made to flow through the circuit at the point 7 2 (corresponding to 
drum rotation) with little or no, or even reversed, voltage supplied 



364 



E. W. KELLOGG 



[j. a M. P. E. 



by EI (or by way of the sprocket). If we imagine, further, that the 
condenser C 2 has the property of showing greater and greater 
capacity as the d-c. voltage across it becomes less, we can see that 
the auxiliary source of power 2 may serve greatly to improve 
the filtering, particularly if the movable idler is not employed. 




FIG. 21. Film path in first commercial magnetic drive recorder, 
type PR-4. 

Some experiments were tried with a rubber bellows or sylphon, 
52, with an air vent instead of an oil dash-pot. There is some diffi- 
culty in making a small metal sylphon sufficiently flexible for use in 
an air-damping system. The air-damper combines the properties 
of the oil dash-pot and the intervening spring in that it permits small, 
quick movements while providing strong damping forces for dis- 
turbances of lower frequency. 

Fig. 19 is a front view of the original magnetic drive recorder. 



April, 1937] 



THE QUEST FOR CONSTANT SPEED 



365 



At the time this photograph was taken the machine was no longer 
in its youth, and its looks have suffered as a result of its having had 
64 teeth extracted. In other words, the buffer sprocket had been re- 
moved. 

Fig. 20 shows the damping magnet, slip-rings, the box containing 
the gear-train, and the driving motor. The motor was mounted 
upon cushions, and drove the recorder through flexible couplings. 
The performance of this machine was excellent, and sufficed to es- 




FIG. 22. Type PR-4 recorder, showing film loops. 

tablish the value of the magnetic drive. The photographs cor- 
roborate the statement that reliance for good performance was placed 
upon the design of the filtering system, and not upon precision work- 
manship. 

First Commercial Magnetic Drive Recorder Type PR-4. The film 
|loop flexibility that it was possible to obtain by means of the mag- 
netic drive so far exceeded expectations that the flexibly mounted 
'idler and the buffer sprocket were found to be superfluous. Ex- 
perience with guiding the film led to the conclusion that more satis- 
factory guiding can be provided when the film is not too slack where 



366 E. W. KELLOGG [J. S. M. p. E. 

it is fed onto the drum but practically free from tension where it 
leaves the drum. For this reason, in subsequent machines employing 
the magnetic drive, the magnet current is given such a value that 
the slight residual tension on the film is such as to pull back on the 
drum, rather than to assist rotation. Fig. 21 shows the film path in 
the recording head. A large amount of wrap around the guide 
roller 6 is provided. Rollers 5 and 7 are both movable, so as to pro- 




FIG. 23. Rear view of PR-4 recorder. 

vide for holding the film in close contact with the guide-roller. The 
arrangement for supplying power to the lower pressure-roller was 
unnecessary, and there is sufficient tension on the film, due not only 
to the tension of the loop between rollers 2 and 5, but also to the 
bearing friction of rollers 5, 6, and 7, to insure the film's maintaining 
good contact with the drum. Roller 9 is so placed as to cause the 
film to be under a very slight tension in the right-hand loop, which 
was regarded as desirable in the way of insuring against the film's 
failing to press tightly against the drum. It is quite possible for a 



April, 1937] 



THE QUEST FOR CONSTANT SPEED 



367 



film loop to resist having film fed to it, or in other words to exert a 
force in the opposite direction from tension. Figs. 22 and 23 are 
photographs of the PR-4 recorder. More recent recorders have 
differed from this primarily in the gear-box arrangement. 

One further comment on the magnetic drive might not be amiss. 
It is, of course, desirable that the drum shaft and bearings shall be in 
good shape, but it is important not to confuse smoothness of action 




FIG. 24. Rotary stabilizer construction of F. J. Loomis and E. W. Reynolds. 

with minimum bearing resistance. For example, a light oil may 
give less total bearing drag than a heavier oil, but the heavy oil 
maintains a more nearly complete film and reduces the amount of 
actual contact or rubbing of surfaces. All contact friction in the 
bearings is somewhat irregular, whereas the oil viscosity effect is 
smooth in action. In using the heavier oil we substitute viscous 
drag (which does no harm and may even be helpful for its contribution 
to the damping) for rubbing friction, which is objectionable. The 
extra power required is easily obtained by an increase in magnet cur- 
rent. Operating with higher magnet current gives more rapid ac- 



368 E. W. KELLOGG [J. S. M. p. E. 

celeration of the flywheel, and more rapid damping out of the tran- 
sient oscillation incidental to starting; but in view of the require- 
ment of low film loop tension, the high magnet current is permis- 
sible only if there is enough bearing resistance to justify it. Light 
oil is desirable for testing bearings, but heavier oil, even though the 
flywheel may not spin as easily, may be better for operation. 

The Rotary Stabilizer. Simpler arrangements for securing the ad- 
vantages of the magnetic drive have been the objective of develop- 
ment engineers, and a happy outcome of some of this work has been 
the "rotary stabilizer." It has been widely used in sound heads and 
has also been used in portable recorders and optical printers. 

Fundamentally, the principle is the same as that of the magnetic 
drive, in that more or less moment of inertia is provided on the drum 
shaft, the drum speed is controlled by tension on the film, and the 
necessary damping is provided by viscous* coupling of the drum 
shaft to a coaxial member rotating at the same or nearly the same 
speed. The essential requirement for damping is that the connection 
be viscous in the sense that the force must increase with the increase 
of relative velocity. 25 This is true of the force imparted through an 
oil film as well as of the reaction between a magnet and a conductor. 
Ordinary rubbing friction (without at least some effect of a viscous 
lubricant) will not provide damping, and for satisfactory results 
friction of the rubbing type must be practically absent. 

In some of the early experiments with damping by coupling the 
flywheel on the drum shaft with a coaxial flywheel, without auxiliary 
drive, the damping was found to be inadequate. C. R. Hanna, of 
the Westinghouse Electric and Manufacturing Co., showed that with 
such an arrangement critical damping could not occur unless the 
viscously coupled flywheel possessed at least 8 times the moment of 
inertia of the member solidly connected to the drum shaft. 26 Better 
proportioning of the moments of inertia therefore solved the problem 
of providing sufficient damping. 

Fig. 24 shows the rotary stabilizer as developed by F. J. Loomis 
and E. W. Reynolds. 27 - 28 In the machines that have employed the 
rotary stabilizer, no movable idlers have been employed to increase 
the flexibility of the film loop. It is therefore essential that the 
film tension be low, in order that loop flexibility shall be adequate. 
There is no auxiliary drive, and the film must supply the torque 

* Eddy current drag and viscous resistance are functionally equivalent (see 
reference 24. p. 660; and U. S. Patents Nos. 1.892,554 Re 19,270 and 1,969,755). 



April, 1937] THE QUEST FOR CONSTANT SPEED 369 

necessary to overcome bearing friction. The success of the system 
thus depends upon minimizing the friction load. The mere expedient 
of employing ball bearings will not insure low friction, but with suit- 
ably selected and properly mounted ball bearings, the friction load 
can be made very low. A single roller serves as guide and pressure 
roller, and this is also provided with ball bearings. With this de- 
sign it has been found possible to keep the pulling loop slack enough 
to provide good filtering. 

If there is much bearing friction between the inner flywheel and 
the drum shaft, the two will rotate with no relative motion unless 
the accelerations of the drum shaft exceed a certain value. The 
result is that although the device might be effective for damping out 
oscillations of large amplitude, the damping would disappear as soon 
as the amplitude dies down below this critical value. Tests and 
practical experience have indicated that with the ball bearing mount- 
ing of the inner flywheel that is employed, damping is effective for all 
accelerations exceeding an extremely small magnitude. The ten- 
dency of bearing friction to lock the flywheel and drum shafts to- 
gether is no doubt materially reduced by the fact that the direction 
of the gravity load with respect to the bearing is constantly changing, 
thus tending to break loose any slight binding condition between the 
balls and their races. Mention has been made of the close analogy 
between the rotary stabilizer and the magnetic drive. Fig. 18 may 
be used to represent the electrical analogue of the stabilizer, the only 
change being that the magnet drive elements indicated opposite the 
bracket in Fig. 18 are eliminated and the connection indicated at 5 
is substituted. (The circuit branch that represents the movable 
idler is, of course, omitted.) L\ represents the moment of inertia of 
the drum and shell. R\ represents the resistance of the oil film to rela- 
tive movements, instead of the magnetic drag; and L 2 represents the 
moment of inertia in the inner flywheel. L% is of zero resistance 
since no additional power is required to keep this inner flywheel ro- 
tating. The practical difference between the two systems is that, as 
compared with the magnetic drive, the effective masses or moments of 
inertia of the rotary stabilizer are very much smaller. If the film is to 
get the rotary stabilizer up to speed in a reasonably short time and 
without undue strain, it is not feasible to employ nearly as large 
masses as are represented in the magnetic drive; and since the film 
loop flexibility can hardly be made any greater than that in the mag- 
netic drive, it is not practically possible to give the filter as low a cut- 



370 E. W. KELLOGG [J. S. M. P. E. 

off frequency. In other words, the magnetic drive can go farther in 
filtering out low-frequency disturbances introduced at the sprocket. 
There are also sources of disturbance that operate directly upon the 
drum shaft. One of these is bearing friction and another is irregulari- 
ties in the film, such as splices and kinks. These sources of dis- 
turbance are indicated in Fig. 18 by alternator D 3 in series with LI to 
represent bearing irregularities, and an alternator D2 in series with Cz 
to represent film irregularities. The ability of the system to resist 
speed changes due to these disturbances is directly proportional to 
the mechanical impedance at the drum, and in this respect the 
magnetic drive possesses an advantage. On the other hand, the 
magnetic drive may introduce disturbances if the magnet rotation 
is not constant, but it can readily be shown that this effect is small. 24 

In performance, the rotary stabilizer is regarded as a close second 
to the magnetic drive and it is obviously a much simpler and less 
expensive device. 29 

Sound Printers. It has been the general practice to print sound- 
tracks on a sprocket, both films being wrapped around the sprocket 
and the exposing light being projected through the negative film onto 
raw stock from the interior of the sprocket. By proper choice of 
the sprocket diameter a certain amount of compensation may be pro- 
vided for the fact that, on the average, the negative film will have 
undergone some shrinkage while the raw stock will not. It is in- 
evitable, however, that continual readjustments must take place of 
the positions of the two films with respect to each other and to the 
sprocket. These readjustments are all of very small amplitude, but 
occur with the engagement of each new tooth. The effect of any 
slippage of one film with respect to the other is to blur the print, 
while irregularities in movement of the raw stock result in variations 
of exposure. If the exposing light-beam is made very narrow, the 
blurring tends to become less, while the variations in exposure be- 
come more noticeable. If the light-beam were made as narrow as a 
recording slit, slipping would show its effects not so much in blurring 
as in periodic stretching out and compression of the waves; or, in 
other words, in introduction of true "wows" or speed fluctuations 
that are not in the negative. In practice, the light-beam in sprocket 
type printers has been made of such width that the principal effect is 
periodic blurring. Nevertheless, this effect is the outcome of speed 
variations, and hence is a logical part of this paper. The Bedford 
non-slip printer, 23 in eliminating this fault, has contributed in an im- 



April, 1937] 



THE QUEST FOR CONSTANT SPEED 



371 



portant degree to quality improvement. The printer is based upon 
the same principle that has already been discussed in connection 
with projectors. 22 A printer employing the same principle was in- 
dependently devised by R. V. Wood. 31 

Fig. 25 is a schematic diagram. In the case of the printer, the 
drum itself does not run at fixed speed as in Bedford's projector, but is 
driven by the negative film. The raw stock must accommodate it- 
self to the surface speed of the moving negative that propels the raw 
stock past the printing point. It would be possible to drive the 
drum and cause the curvature of both films to change in accordance 
with shrinkage, but the expedient employed by Bedford is simpler. 




To Hold-Back 

Sprocket 



FIG. 25. Film paths in Bedford non-slip printer. 



If the drum failed to run at uniform speed, the only effect upon 
the print would be variations in exposure. A rotary stabilizer is 
mounted on the drum shaft to minimize speed fluctuations. The 
printing is done by a concentrated beam of light so that it all occurs 
where the films are in tight contact. A degree of contact is possible 
in this printer that could not be permitted in the sprocket type, for 
the reason that continual readjustment to the sprocket teeth must 
be permitted in the latter, or the teeth will begin to cut the film at 
the perforations. 

Fig. 26 shows one of the laboratory models of the non-slip printer. 
A number of commercial designs have been built and are now in use 
in several laboratories. 



372 



E. W. KELLOGG 



[J. S. M. p. E. 



Gates. It is frequently convenient to support the film where it 
passes the scanning light by means of a stationary support over which 
the film slides. This necessarily means more or less sacrifice of 
speed constancy, since the speed constancy is best where the film is 
held against the drum, and any fluttering of unsupported film or any 
inequalities in shrinkage between the driving and the scanning points 
introduces variations in velocity. The stationary support or "gate," 
as it is commonly called, also introduces friction. 




FIG. 26. Laboratory model of non-slip printer. 



In many gate type reproducers, this friction has been unnecessarily 
large. It can be made extremely small. If in addition to minimiz- 
ing friction, the design is such as also to minimize the amount of 
unsupported film as well as the total length of film between the drum 
and the scanning point, the harmful effects of employing a gate can 
be made quite small. 



April, 1937] THE QUEST FOR CONSTANT SPEED 373 

Fig. 27 shows the machine of the gate type that has come the closest 
of any within the author's acquaintance to matching the performance 
of a machine in which the scanning is done directly on a drum. 
Motors. Nothing has been said so far about the electric motor, upon 
lich ultimate dependence is placed for speed constancy. For- 




FIG. 27. Laboratory film phonograph employing "gate." 

tunately, the larger power supply systems provide a very high order 
of constant frequency, and the driving motors are electrically coupled 
to generators in which many tons of steel and copper are moving at a 
velocity of the order of 100 feet per second. Sudden frequency 
changes, such as to impair sound quality, do not often occur under 
such conditions. Small synchronous motors have been developed 



374 E. W. KELLOGG [J. S. M. P. E. 

that have given no trouble with respect to "hunting." The type of 
construction is such as to provide very effective pole-face grids. 

Various schemes have been worked out to meet the condition where 
the power supply is direct current. Centrifugal type governors, 
operating generally to control the field current, have been developed, 
which are quite satisfactory. It is, of course, necessary to avoid 
the type of governor that causes the speed to fluctuate above or 
below a correct average by alternately opening and closing a con- 
tact that short-circuits a part of the field resistance. If, however, 
this opening and closing occurs at an extremely rapid rate it is per- 
missible. A number of systems have been worked out that regulate 
the motor in response to changes of frequency in a tone generated by 
a small high-frequency alternator coupled to the motor. Such a 
system was described by H. M. Stoller, 1928. 30 

Where several machines are operated synchronously, Selsyn motors 
are often employed. This arrangement practically gears the ma- 
chines together, so that they can be brought up to speed without any 
relative phase-shifts. There are a few precautions that must be ob- 
served to insure that there will not be any slipping of poles during 
the starting, but with proper installation, the Selsyn systems have 
given very little trouble and do not seem to have been responsible 
for any appreciable speed fluctuations during operation. Although 
the connection between Selsyn machines is an elastic coupling, the 
damping is of the same order of magnitude as that of synchronous 
motors of comparable rating, and little trouble from hunting has 
been encountered. This assumes polyphase connections of both 
primary and secondary windings. 

Other Designs and Inventions. One of the purposes of this paper 
has been to call attention to some of the ingenious devices that have 
been employed or suggested for improving speed constancy, and to 
give credit to those who have taken part in this "Quest." Among 
the workers in other countries, Messrs. Vogt, Massole, and Engl de- 
serve mention. Obviously, no review of this kind can include all 
the arrangements and inventions that deserve mention, and the 
list given here is necessarily limited to devices that have come to the 
writer's attention. It is to be hoped that other writers, perhaps 
through contributed discussions or in separate papers, will supple- 
ment this outline by describing other interesting arrangements and 
schemes that ought to have been included, especially if the devices 
in question have not been described in earlier technical publications. 



April, 1937] THE QUEST FOR CONSTANT SPEED 375 

REFERENCES 

1 BELL, A. G., AND TAINTER, S.: U. S. Pat. No. 341,214. 

J MAXFIELD, J. P., AND HARRISON, H. C.: "High-Quality Recording and 
Reproducing Music and Speech," Trans. A. I. E. E., 45 (Feb., 1926), No. 2, 
p. 334. 

FREDERICK, H. A.: "Recent Advances in Wax Recording," Trans. Soc. 
Mot. Pict. Eng., XH (1928), No. 35, p. 709. 

4 KELLOGG, E. W.: "Electrical Reproduction from Phonograph Records," 
Trans. A. I.E. E., 46 (June, 1927), No. 6, p. 903. 

6 RICE, C. W., AND KELLOGG, E. W.: "Notes on the Development of New 

of Hornless Loud Speaker," Trans. A. I. E. E., 44 (April, 1925), No. 4, 
>. 982. 

BLATTNER, D. G., AND BOSTWICK, L. G.: "Loud Speakers for Use in Thea- 
ters," /. Soc. Mot. Pict. Eng., XIV (Feb., 1930), No. 2, p. 161. 

WENTE, E. C., AND THURAS, A. L.: "A High-Efficiency Receiver for Horn 
Type Loud Speakers," Bell Syst. Tech. J., VH (Jan., 1928), No. 1, p. 140. 

7 MALTER, L.: "Loud Speakers and Theater Sound Reproduction," /. Soc. 
Mot. Pict. Eng., XIV (June, 1930), No. 6, p. 611. 

OLSON, H. F. : "Recent Developments in Theater Loud Speakers of the 
Directional Baffle Type," J. Soc. Mot. Pict. Eng., XVIII (May, 1932), No. 5, 
>. 571. 

8 FREDERICK, H. A.: "Vertical Sound Records: Recent Fundamental Ad- 
vances in Mechanical Records on Wax," /. Soc. Mot. Pict. Eng., XVIII (Feb., 
1932), No. 2, p. 141. 

BARTON, F. C.: "Victrolac Records," J. Soc. Mot. Pict. Eng., XVIH (April, 
1932), No. 4, p. 452. 

"High Fidelity Lateral Cut Disk Records," J. Soc. Mot. Pict. Eng., XXII 
(March, 1934), No. 3, p. 179. 

DIMMICK, G. L., AND BELAR, H.: "Extension of the Frequency Range in 
Film Recording and Reproduction," J. Soc. Mot. Pict. Eng., XIX (Nov., 1932), 
No. 5, p. 401. 

10 CONSTABLE, J.: U. S. Pat. No. 1,425,177. 

11 AMET, E. H.: U. S. Pat. No. 1,221,407. 

ia ELMER, L. A., AND BLATTNER, D. G.: "Machine for Cutting Master Disk 
Records," Trans. Soc. Mot. Pict. Eng., XIII (1929), No. 37, p. 227. 

13 BEDFORD, A. V.: "A Planetary Reduction Gear System for Recording 
Turntables," /. Acoust. Soc. Amer., HI (Oct., 1931), No. 10, p. 207. 

U. S. Pat. No. 2,027,666. 

14 MEAD, M. S.: U. S. Pat. No. 1,854,949. 

16 PPANNENSTIEHL, H.: "A Reproducing Machine for Picture and Sound," 
Trans. Soc. Mot. Pict. Eng., XIH (1929), No. 38, p. 253. 
U. S. Pat. No. 1,922,669. 

16 KELLOGG, E. W., AND BELAR, H. : "Analysis of the Distortion Resulting 
from Sprocket-Hole Modulation," /. Soc. Mot. Pict. Eng., XXV (Dec., 1935), 
No. 6, p. 492. 

17 SHEA, T. E., MACNAIR, W. A., AND SUBRIZI, V. : "Flutter in Sound Records," 
/. Soc. Mot. Pict. Eng., XXV (Nov., 1935), No. 5, p. 403. 

18 HOXTB, C. A.: U. S. Pat. No. 1,758,794. 



376 E. W. KELLOGG 

19 HEISLER, C. L.: U. S. Pat. No. 1,808,551. 

20 HEISLER, C. L.: U. S. Pat. No. 1,831,562. 

21 HOXIE, C. A.: U. S. Pat. No. 1,756,864. 

22 BEDFORD, A. V.: U. S. Pat. No. 1,754,187. 

23 BATSEL, C. N.: "A Non-Slip Sound Printer," J. Soc. Mot. Pict. Eng., 
XXIII (Aug., 1934), No. 2, p. 100. 

24 KELLOGG, E. W.: "A New Recorder for Variable-Area Recording," J. Soc. 
Mot. Pict. Eng., XV (Nov., 1930), No. 5, p. 653. (The title of this paper is mis- 
leading, in that the main features of the recorder were in no wise related to the 
type of sound-track to be recorded.) 

24 ROBINSON, L. T.: U. S. Pat. No. 1,880,106. 

24 HANNA, C. R.: U. S. Pat. No. 2,003,048. 

27 REYNOLDS, E. W.: U. S. Pat. No. 2,013,109. 
LOOMIS, F. J.: U. S. Pat. No. 2,019,147. 

24 LOOMIS, F. J., AND REYNOLDS, E. W. : "A New High-Fidelity Sound Head," 
J. Soc. Mot. Pict. Eng., XXV (Nov., 1935), No. 5, p. 449. 

29 COOK, E. D.: "The Technical Aspects of the High-Fidelity Reproducer," 
/. Soc. Mot. Pict. Eng., XXV (Oct., 1935), No. 4, p. 289. 

STOLLER, H. M.: "Synchronization, and Speed Control of Synchronized 
Sound Pictures," Trans. Soc. Mot. Pict. Eng., XII (1928), No. 35, p. 696. 

31 WOOD, R. V.: "A Shrinkage Compensating Sound Printer," /. Soc. Mot. 
Pict. Eng., XVin (June, 1932), No. 6, p. 788. 



THE EFFECT OF ABERRATIONS UPON IMAGE QUALITY* 
W. B. RAYTON AND A. A. COOK** 



Summary. Lenses are used to form images for two principal purposes 
first, to produce the most accurate record possible of the original object; and second, 
to produce a pleasing effect. The character of the image formed by a lens depends 
upon diffraction and upon the residual aberrations remaining after the designer 
and the manufacturer have done their best. For pictures of the first type it is de- 
sirable that aberrations be reduced to a minimum, but for pictures of the second type 
they are very often deliberately employed to produce desired effects. In motion pic- 
ture projection, lenses of the first class are doubtless always desired. In motion pic- 
ture photography, some attention has been given to achieving special effects by de- 
liberately introducing aberrations into the lens. 

Among the many aberrations that afflict lenses, one of the most important is chro- 
matic. Since, in general, only two colors can be brought to a common focus, some 
thought has been given to the question of what two colors it is best to choose to meet the 
requirements of various kinds of lighting and different types of sensitivity of the emul- 
sion. Recent experiments indicate that for a combination of particular interest in 
motion picture photography, namely, incandescent lighting and super-pan emulsion, 
no significant difference in performance is detectable among lenses of 12-inch focus 
or less, depending upon whether the two colors chosen for chromatism are yellow and 
violet, or red and violet. 

The discussion that follows will be confined to photographic and 
projection lenses. The function of such lenses is to form images, 
and generally images in which each detail in the object is reproduced 
with the utmost fidelity. Any object whatsoever may be regarded 
as consisting of a vast number of points, and in the ideal image each 
point of the object would appear as a point in the image. The 
degree of approach to perfection in an image, or, if one prefers, the 
degree of departure from perfection, can therefore be studied by in- 
vestigating the character of the images of several judiciously chosen 
object points. 

No lens can possibly form a point-image of a point object. We 
were told in elementary physics that light travels in straight lines; 

* Received February 12, 1937; presented at the Fall, 1936, Meeting at 
Rochester, N. Y. 
** Scientific Bureau, Bausch & Lomb Optical Co., Rochester, N. Y. 

377 



378 W. B. RAYTON AND A. A. COOK fj. S. M. P. E. 

but because of its wave nature, light refuses to behave in this manner 
when it passes the edge of an opaque object. Close examination will 
reveal that the edge of the shadow is not absolutely sharp and that 
some light is apparently bent into the shadow as it passes the edge 
of the opaque object. The same effect occurs at the rim of a lens; 
so that even if there were no other factors responsible for imperfection 
in the image, the light arising in a single object-point, instead of 
uniting in an image-point, would be distributed over an area of 
finite size in the manner illustrated in Fig. 1. The phenomenon 
responsible for this effect is called diffraction. In photographic and 



08 




o a i o 

FIG. 1 . Distribution of light in the diffraction image of a star. 

projection lenses it plays a very small part, and we shall not refer to 
it again. 

Even if we ignore diffraction, however, we can not meet the re- 
quirements of perfect image formation. In the language of geometri- 
cal optics, perfect imagery requires that a lens unite in a common 
image-point all the rays of light that leave the corresponding 
object-point within the range of the angular aperture of the lens. 
But no lens can do this, and as a result, no object-point can be imaged 
as a point but must appear as an area of finite size within which light 
may be distributed in a great variety of ways. Necessarily, it follows 
that if two object-points are too close together their images will over- 
lap so badly that it will be impossible to recognize from a study of the 



April, 1937] 



ABERRATIONS AND IMAGE QUALITY 



379 



image that there are two object points. This means that the image 
can not contain all the detail of the object, outlines are not absolutely 
sharp, light falls into areas that should be black and turns them gray, 
and, in general, the image, we say, is not sharp or does not have 
perfect definition. 

While from the standpoint of performance, the end result is the 
same, regardless of how we name the cause, it has been very useful 

(Lens axis to this side) 

flflDDDB 



o 



100 150 



20 



250 



T* Focal Plane of Anastlgmat 



Within focus 




BDBD 




Rapid Rectilinear Lens 
At 10 Obliquity 

FIG. 2. (Upper.) The image of a star as it appears in the focal plane of 
a good anastigmatic lens. The plate was focused for best definition at the 
center of the field. Successive exposures were at 5-degree intervals from 
the center of the field to a point 25 degrees from the center. 

(Lower.) A series of exposures made with a rapid rectilinear lens at a 
field of 10 degrees, beginning with an exposure in the focal plane that shows 
the image to consist of practically a straight horizontal line and followed by 
six other exposures, each plate being moved 1 mm. nearer the lens than its 
predecessor. These figures show the combination of coma and astigmatism. 

to analyze this departure from perfection into separate entities called 
aberrations. Some of these, called "chromatic" aberrations, are 
due to the different refractivities of glass for light of different wave- 
lengths or colors. Another called "spherical" aberration, is due to the 
finite aperture of the lenses; it increases in amount with increase of 
aperture. Other aberrations appear only in the case of light passing 
obliquely through the lens, as must occur in imaging any object- 
point outside the center of the field. Fig. 2 serves to demonstrate 
some of the irregular patches of light that must serve as the best 
attainable images of object-points under various conditions. It is 



380 



W. B. RAYTON AND A. A. COOK [J. S. M. P. E. 




* LTI 




In a lens with considerable spherical aberration there are fre- 
quently two points at different distances from the lens at which 
some approach to a sharp image is attainable; between these 
positions the image is inferior to either of the two best positions. 

Figs. 4 and 5 show the two positions in which some semblance of 
image quality is obtained; Fig. 6, which is definitely inferior, is 
taken in a plane lying between the other two. 



April, 1937] ABERRATIONS AND IMAGE QUALITY 381 

not the present purpose to discuss the nature of these aberrations at 
any length, since they have been considered in several papers pre- 
sented to the Society in the past. l It is more the purpose to discuss 
some of the effects of these aberrations upon image quality and 
some of the dilemmas faced by the lens designer in his effort to reduce 
them to a minimum. 

It is evident from an inspection of the photographed images shown 
in Fig. 2 that the distribution of light within these areas is far from 
uniform. This leads at once to the result that the definition of a 
photographic lens will appear to vary with exposure and development. 
Underexposure and high-contrast development will fail to reveal 



FIG. 3. Special aberration curves. For perfect correc- 
tion the curve of aberration should be a straight vertical 
line, (a) Typical under-correction; (b) slight over-cor- 
rection at the margin, and an under-corrected zone; (c) 
shows slight under-correction for the margin and an over- 
corrected zone. 

aberration effects that, on the other hand, become overemphasized 
by overexposure and low-contrast development. Lenses can differ 
over a wide range both as to the character and the amount of spherical 
aberration present. In a simple lens the focal length decreases 
progressively for each zone as the distance of the ray from the center 
of the lens increases. Such a condition is indicated by the curve a 
in Fig. 3. This represents pure under-correction. Curve b in the 
same figure shows a condition common in photographic and projection 
lenses wherein correction is complete for some one zone, while there 
exists over-correction for zones farther from the center and under- 
correction in zones nearer the center. Such a condition is responsible 
for what is known as stop-difference. If the lens is focused with full 
aperture and then stopped down, readjustment of focus is necessary 



382 



W. B. RAYTON AND A. A. COOK 



[J. S. M. P. E. 



to obtain the best definition. Still another condition is represented 
in curve c. It seems obvious that such wide difference in the charac- 
ter of the residual spherical aberration should lead to substantial 
differences in the distribution of light in the image. In some cases 
it leads apparently to two image-planes at different distances from 
the lens. Figs. 4 and 5 show the character of the images in two such 
image-planes. Fig. 6 is a photograph of the image in a plane inter- 
mediate between these other positions. 

From such considerations it is very obvious that different operators 




FIG. 7. Plate holder used in chromatic aberration tests. 

may produce very different results with the same lens without under- 
standing why they differ, while a skillful operator can at will use the 
same lens to produce quite different effects. 

As a matter of fact, the effort has been frequently made by the 
designer to produce lenses that, by virtue of aberrations deliberately 
left in the lens, produce images characterized by a distinct haziness. 
Such lenses are popular for portraiture in as much as they have a ten- 
dency to suppress small skin blemishes that would otherwise have 
to be removed from the negative by a retoucher, and because there is 
a general belief that such "soft-focus" pictures are more life-like 
representations of the features of the subject than are pictures made 
with a sharp-focus lens. There is room here for nice judgment as to 



April, 1937] ABERRATIONS AND IMAGE QUALITY 383 

just how much aberration to leave in the lens; hence the opportunity 
for the existence of many lenses of this general character. In some 
cases the problem has been solved by making the amount of aberra- 
tion, and hence the degree of diffusion, variable at the will of the 
operator by providing an adjustment that permits varying the posi- 
tion of one of the component elements of the system. 

There is a further choice for the designer to make in deciding what 
kind of aberration to employ for producing the soft-focus effect. 
He can choose spherical aberration or chromatic aberration, or both. 
Since in the presence of chromatic aberration there are many focal 
points each corresponding to a different color, many different effects 
can be produced, depending upon the character of the emulsion used. 
In focusing the camera, the eye will select the image formed by 
green light if the subject is illuminated with anything approaching 
white light. The emulsion may have its sensitivity donfined to the 
blue, in which case the image will look simply like an out-of -focus 
image when the photograph is made. On the other hand, if an 
emulsion is used that is almost equally sensitive to a long range of 
colors, a truly diffuse image will result, but one characterized by 
more unpleasant lack of sharpness of outline than in soft-focus lenses 
that depend upon spherical aberration for the diffusion. 

The reference to chromatic aberration leads naturally to considera- 
tion of another matter that has given rise to considerable speculation 
in optical circles, viz., the most profitable type of chromatic correction 
to employ. This has been discussed for the case of telescope objec- 
tives 2 for both visual observation and for photographic use, as well 
as for camera lenses. 3 In general, the designer can bring to a common 
focus two colors only, and that only for a single zone of the lens. 
Other colors will come to foci nearer or farther from the lens. 

It has been standard practice to achromatize visual instruments 
for colors corresponding to the C and F lines of the spectrum. The 
C line is orange-red in color, and the F is blue. Photographic lenses 
have usually been corrected for the D line, which is yellow, and the 
G' line, which is violet. The wisdom of this choice in both cases 
has been called into question, and particularly in the case of camera 
lenses was it questioned upon the introduction of supersensitive 
panchromatic film and tungsten lights into motion picture practice. 
There has been some argument in favor of achromatizing for the C and 
G f lines for use under these conditions. The writers made some 
rather hasty tests in 1928 employing lenses that had been corrected 



384 W. B. RAYTON AND A. A. COOK [J. s. M. p. E. 

for the yellow and the violet in the classical manner, with different 
kinds of illuminations and emulsions, to see whether there was any 
perceptible degradation of image quality resulting from the use of 
tungsten light and superpan emulsions, as compared with the per- 
formance of the same lenses used with orthochromatic emulsions 
and arc lights. No differences were discovered. These experiments 
were reported to a joint meeting of the Society and the Academy of 
Motion Picture Arts &* Sciences. More recently we have investi- 
gated the matter somewhat more thoroughly by means of an experi- 
ment that it may be of interest to describe and with results that will 
be set forth shortly. 

For this experiment we made three special lenses, each of 320-mm. 
focal length (about I2 l /s inches), with a relative aperture of //6. One 
of them was achromatized for the C and F lines, or red and blue ; an- 
other for the Sfand G f , or yellow and violet, corresponding to the 
classical practice in photographic lens design; and the third for C 
and ', or red and violet. 

With each of these three objectives test photographs were made 
on 2 X 5-inch plates using process, orthochromatic, and supersensi- 
tive panchromatic emulsions. As a test-object we used a series of 
parallel vertical lines engraved upon glass and stationed in the focal 
plane of a collimator whose objective lens had a focal length of 1800 
mm. The object appeared then to be at infinity. Because of the 
difference in the focal length of the collimator lens and the lens under 
test, the chromatic aberrations of the collimator lens were reduced 
to an utterly insignificant value in the focal plane of the lenses under 
test; namely, to about one-thirtieth of their actual value. A plate 
holder specially made for the purpose, shown in Fig. 7, supported 
the plate at an angle of 2 degrees to the optical axis of the system. 
Spring clips held the emulsion surface firmly against a shoulder. A 
device was provided to permit a fine needle to be drawn across the 
center line of the plate to mark the position of the best visual focus. 
The needle holder had a set-screw and lock-nut adjustment to fix its 
position so that it could move only in a straight line at right angles 
to the optical axis of the camera. The procedure employed in making 
a test negative was as follows: 

The lens was mounted upon the camera front board and the plate 
holder, unloaded, was put into position. Observing through a 25X 
focusing microscope, the plate holder was first racked back into 
position until the needle point was in the focal plane of the micro- 



April, 1937] 



ABERRATIONS AND IMAGE QUALITY 



385 



scope; then the target image formed by the lens was brought to the 
same plane by focusing the lens. This method was found to be 
accurate and reproducible, because there is parallax between the 
needle point and the image whenever they are not exactly in the same 
plane. 

The camera front and back plates were then locked; the plate 



FIG. 8. Reproduction, ten times enlarged, of one of the photographs 
obtained in the chromatic aberration tests. The horizontal line is the plane 
of best visual focus. Distances above this line are farther from the lens. 
The best photographic focus in this particular case is about 0.5 mm. farther 
from the lens than the visual focus. The three exposures in the group at 
the left were made with tungsten light, and the three at the right with 
daylight. 

was loaded into the holder, a line scratched across its center with the 
needle fixture, and the holder put back into the camera for a series 
of exposures. On each plate three exposures were made with tungs- 
ten light and three with daylight illuminating the target. The 
plate holder was arranged to move laterally between exposures. 
Fig. 8 shows the kind of result obtained. This is a WX enlargement 
from the original negative. 

The horizontal line in the figure indicates the position of best 
visual focus. Everything above this line is nearer the lens and 



386 



W. B. RAYTON AND A. A. COOK 



[J. S. M. P. E. 



everything below is farther away. The area through which the 
target lines show up sharply represents the depth of focus of the lens 
under test. It amounts to just about 1.0 mm., corresponding to a 
depth of field that extends from 335 feet to infinity. The center of 
that region of sharpness was assumed to be the point of best photo- 
graphic focus, and the measurements shown in Table I were made 
between this point and the line marking the visual focal point. The 
measurements were made with a linear comparator on the original 
negatives. Fig. 8 is a reproduction of the negative made with the 
lens that was corrected for the red and blue on a process plate 
sensitive only to blue and violet light. The three exposures on the 
left were made with tungsten light and the other three with daylight. 
Similar exposures were made with the same lens on an orthochromatic 
plate and on a supersensitive panchromatic emulsion, all at the same 
visual focus setting. It is apparent by inspection of Fig. 8 and the 
data in Table I that with all three types of emulsion the photographic 
focus of this lens is farther from the lens than the visual focus. 



TABLE I 





Emulsion 


Color Correc- 






Supersensitive 


tion of Lens 


Process 


Orthochromatic 


Panchromatic 


C-F 


Tungsten +0.93 mm 


+0.57 mm. 


+0.18 mm. 




Daylight +1.07 


+0.78 


+0.26 


D-G' 


Tungsten +0.21 


+0.07 


+0.22 




Daylight +0.26 


+0.14 


+0.14 


C-G' 


Tungsten +0.33 


+0.12 


+0.20 




Daylight +0.38 


+0.23 


+0.17 



These values are displacements of best photographic focus from best visual 
focus. Each value is the average of several determinations. Plus values indi- 
cate that the best photographic focus is farther from the lens than the best visual 
focus. 



A similar series of tests with both kinds of illumination and the 
three emulsions was made with each of the other two lenses. The 
photographs are not reproduced here but the results of the measure- 
ments made on the plates are set forth in Table I. Under C-F for 
example, the figure is +0.93 mm. for tungsten illumination. This 
means that the average photographic focus was 0.93 mm. farther 
from the lens than the visual focus, determined as an average of 12 



April, 1937] ABERRATIONS AND IMAGE QUALITY 387 

separate readings, the largest of which was 1 mm. and the smallest 
0.82 mm. 

We found that results could be repeated with considerable accu- 
racy, and the probable error of any of the averages is about 0.04 mm. 
This includes all possible sources of error. It is to be noted that 
there is no significant difference between the lens corrected for the 
C and G' lines and the lens corrected for D and G' in respect to their 
behaviors with either orthochromatic or panchromatic materials 
The results of this experiment confirm our earlier tests, indicating 
that a lens corrected for the D and G' lines of the spectrum is at 
least as good as any of the other combinations that have been sug- 
gested even for specific use with tungsten light and panchromatic 
materials. 

These tests were conducted with complete impartiality and with 
the sole purpose of trying to determine by carefully executed experi- 
ments the best type of chromatic correction for photographic lenses 
intended for professional motion picture photography. For the 
designer, one type of correction is no more difficult to achieve than 
another; hence, there was not the slightest incentive for our judg- 
ment to be biased one way or the other. These results should be 
reliable unless there was some error of principle involved that we have 
not been able to detect. 

It is realized that the substance of this paper is not an adequate 
discussion of the subject represented by the title. It will have 
served its purpose, however, if it gives to the reader who is utterly 
unfamiliar with the problems of the lens designer some impression 
of the character of his task. 

It is a pleasure to acknowledge our indebtedness to Professor 
Rudolph Kingslake of the Institute of Optics of the University of 
Rochester, for the use of Figs. 1 and 2. 

REFERENCES 

1 CORY, A. S. : "Optical Requirements of Motion Picture Projection Objec- 
tives," Trans. Soc. Mot. Pict. Eng. (1918), No. 6, p. 9. 

2 SMITH, T. T.: "Color Correction of an Achromatic Doublet," J. Opt. Soc. 
Amer. and Rev. Sci. Instr., 10 (1925), No. 1, p. 39; 15 (1927), No. 5, p. 247. 

1 DUBRAY, J. A.: "Chemical Focus in Cinematography," Amer. Cinemat., 
15 (Oct., 1934), No. 6, p. 248. 



THE ANALYSIS AND SPECIFICATION OF COLOR 5 
K. S. GIBSON** 



Summary. A brief resume of the various methods used to describe, analyze, or 
specify colors, including color names, systems of material color standards, colorime- 
ters, and spectrophotometric methods. The computation of colorimetric quantities 
from spectrophotometric data is considered, together with methods of specification 
based thereon. 

(I) Introduction. 
(II) Color names. 

(III) Color systems. 

(1} Maerz and Paul Dictionary of Color. 

(2) Munsell Book of Color. 

(3) Lovibond Tintometer glasses. 

(IV) Colorimeters. 

(1) Dependent upon material color standards. 
(2} Using spectrum primaries. 
(3) Filter photometers. 
(V) Spectrophotometric analysis. 
(VI) Colorimetric computations. 
(VII) Colorimetric specifications. 
(VIII) Color of illuminants. 
( IX ) Conclusion . 

(I) INTRODUCTION 

Color is an important factor in nearly every article of commerce, 
affecting the sales value of the article in one of three ways : 

(1) The color may be the primary property because of which the 
article is desired. This is true of dyes, pigments and paints, and 
colored glassware used in railway, traffic, and aviation signalling. 

(2) The color may be an indication of the quality or identity of 
an article, as in the case of vegetable and mineral oils, sugars and 
syrups, dairy products, flour, glass, paper, and raw materials such 
as cotton, wool, and silk. 

* Received January 29, 1937; publication approved by the Director of the 
National Bureau of Standards of the U. S. Department of Commerce, Washing- 
ton, D. C. 

** Chief, Colorimetry Section, National Bureau of Standards, Washington, 
D. C. 
388 



ANALYSIS AND SPECIFICATION OF COLOR 389 

(3) The article may be artificially colored for the purpose of pro- 
ducing and increasing sales appeal. Such materials include textiles 
and wearing apparel of all kinds, automobiles, furniture, enamels 
and porcelains, and candy and other food-stuffs. Containers of food 
and other merchandise are often artistically colored for the same 
purpose. 

The annual trade value of these materials is in the billions of 
dollars. The analysis and control of colors is thus seen to be a matter 
of immense importance in commerce and industry. If at the same 
time it be realized that there are millions of perceptibly differing 
colors, any of which it may be desired to analyze and specify, the 
importance of colorimetric research and testing can readily be under- 
stood. 

Various methods of describing colors and of measuring and specify- 
ing color stimuli are available, the most suitable depending upon the 
problem. Most of these methods are briefly described or noted in 
the present paper, although particular instruments illustrating the 
methods are listed in but a few cases. The scope of the present paper 
is limited since it is not desired to duplicate any of the following pub- 
lications : 

(1) The monographs on color recently published by The Inter- 
national Printing Ink Corporation. There are three of these, en- 
titled, Color Chemistry, Color as Light, and Color in Use, all with chro- 
matic illustrations. They introduce the subject of color from the 
point of view of the chemist, the physicist, and the psychologist, 
respectively. 

(2) The Handbook of Colorimetry, 1 prepared under the direction of 
Prof. A. C. Hardy. This publication should be very useful in the 
colorimetry of materials by way of spectrophotometric analysis. 

(3) The colorimetry report being prepared by the Colorimetry 
Committee* of The Optical Society of America, Dr. L. A. Jones, 
Chairman. The report should be an authoritative consensus of 
opinion regarding the nomenclature and technic of colorimetry. Its 
publication is expected in the near future.** 

* This committee includes fifteen or more persons vitally interested in colori- 
metric matters and representative of industrial laboratories, universities, and gov- 
ernment departments. 

** Earlier publications that may profitably be read by those interested include 
a. previous O.S.A. Colorimetry Committee report (L. T. Troland, Chairman),* 
and Color and Its Applications, by Luckiesch.* 



390 K. S. GIBSON [j. s. M. p. E. 

The present paper at most is merely an introduction to the subject 
of colorimetry. To acquire a knowledge of the science adequate to 
enable one to avoid the pitfalls of unsound colorimetric methods and 
to make of colorimetry a safe and useful tool in the analysis, control, 
and specification of the colors of materials requires extensive study 
and experience. Those who wish to undertake any phase of colori- 
metric research or testing should at the least become familiar with 
the subject matter of the publications referred to herein. He will 
then be prepared to pursue the subject along his own special lines, 
both in reviewing the voluminous literature available and in conduct- 
ing the investigations in which he is interested. 

(II) COLOR NAMES 

The simplest way of describing colors is, of course, by name. If, 
for example, one wished to use ten differently colored papers or cards 
for some sort of classification or filing system, he could specify red 
yellow, green, blue, purple, white, brown, pink, black, and gray, 
and the dealer would have no difficulty in supplying materials of those 
colors sufficiently differentiated from each other for the purpose. 
There also are a large number of other names, such as ivory, cream, 
lemon yellow, chrome yellow, chrome green, ruby, scarlet, maroon, 
amber, gold, orange, ultramarine, cobalt blue, violet, magenta, and so 
on, which doubtless in every case have given the reader a more or less 
definite mental picture of the color named ; for each name is, by usage 
or derivation, closely associated with some particular color. 

However, it is a matter of common knowledge that if one wishes to 
match the exact "shade" of a given material, it is a very unsafe pro- 
cedure to order the material by color name. And, in general, while 
such names may serve to describe colors adequately for certain pur- 
poses, they rarely serve to specify them with sufficient precision for 
even the simplest colorimetric work.* However, if one has at hand a 

* A system of color names on a scientific basis using a few of the most common 
names qualified by certain adjectives such as light, dark, pale, weak, etc., has 
been formulated by Dr. I. H. Godlove, representing The Intersociety Color 
Council. A further study and revision of this system is being undertaken by Dr. 
D. B. Judd of the National Bureau of Standards and Mr. K. L. Kelly, represent- 
ing the American Pharmaceutical Association. It is planned to use this system 
for naming the hundreds of drugs and chemicals of the U. S. Pharmacopoeia and 
the National Formulary. Correlation is being made with the Munsell color sys- 
tem, and thereby with the I.C.I, system of color specification, described later in 



April, 1937] ANALYSIS AND SPECIFICATION OF COLOR 391 

sample of the desired color, the selection of a sample of closely match- 
ing color is greatly facilitated. This leads immediately to a con- 
sideration of color cards, atlases, dictionaries, etc., which are discussed 
in the following section. 

Before considering these, however, another class of color names 
may be noted, which, while of undoubted use in trade, can hardly be 
considered as on a scientific basis. It will be sufficient to quote the 
following from Maerz and Paul's Dictionary of Color: 

"Challamel's History of Fashion in France states that in the sixteenth century 
for instance, 'French women wore colours, and great was their number from 
Rat-colour to that called widow's joy, or envenomed money, or chimneysweep'! Nor 
was this the only period when such names flourished; succeeding centuries kept 
pace with and even outdid the sixteenth in the invention of these extraordinary 
names for colors. The eighteenth produced in France: Rash Tears, Paris Mud, 
Brand from the Opera, Burnt Opera House, Stifled Sigh, and similar inspirations, 
. . . ; while in England the European Magazine for January, 1783, describes as 
fashionable: Elliott's Red-Hot Bullets and The Smoke of the Camp of St. Roche. 
... If one is inclined to smile at the whimsical names quoted above, as being in 
any sense an indication of the affected or effete mannerisms of an earlier and less 
civilized age, one should remember that today we match them with Folly, Lucky 
Stone, Basketball, Elephant's Breath, Skating, Wireless, Water Sprite. Perhaps 
there is some deep psychological reason for this poetic feeling towards the names for 
colors." 

The following quotation from the same source may also be of in- 
terest, although it may be uncertain in some cases whether the name 
refers to color, material, or clothing: 

"A great sensation was caused at the opera one night by the arrival of a lady 
dressed as follows: Her gown was 'a stifled sigh,' trimmed with 'superfluous re- 
grets,' with a bow at the waist of 'perfect innocence,' ribbons of 'marked atten- 
tion,' and shoes of 'the queen's hair' embroidered in diamonds, with the 'venez-y- 
voir' in emeralds. Her hair was curled in 'sustained sentiments,' a cap of 'assured 
conquest' trimmed with waving feathers and ribbons of 'sunken eye,' a 'cat' or 
palatine of swansdown on her shoulders, of a colour called 'newly arrived people' 
(parvenus), a 'Medicis' arranged 'as befitting,' a 'despair' in opals, and a muff of 
'momentary agitation.' " 

(HI) COLOR SYSTEMS 

There are several elaborate systems of material color standards 

this paper. This means of designating colors is intended to be "sufficiently stand- 
ardized to be acceptable to science, sufficiently broad to be appreciated and usable 
by science, art, and industry, and sufficiently commonplace to be understood, at 
least in a general way, by the whole public." 



392 K. S. GIBSON [J. S. M. P. E. 

available for general use,* in addition to innumerable "color cards" 
of one kind or another. Three of these systems have been selected for 
brief description to illustrate both the advantages and the limitations 
of this method of color specification. 

(1) A Dictionary of Color, 6 by Maerz and Paul, to which reference 
has already been made. According to the authors, "This work is 
primarily intended as a reference for the individual who seeks to re- 
late colors with the names by which they are commonly identified." 
Its pages are said to contain "the most extensive range of colors as 
yet published, together with a list of practically all recorded color 
names in use up to this time in the English language." The authors 
state that "the inks used in this work . . . are reasonably permanent 
to light and the color plates may be freely exposed." The following 
brief description of the charts is taken from NBS letter circular 
LC-481 : 4 

"Contains approximately 7000 different color samples printed on semi-glossy 
paper, about 6000 of which are Vs by 6 / 8 -inch rectangles, and about 1000 of the 
darker of which are P/ie by 6 /s-inch rectangles. The hue circle is covered in 8 
intervals, each interval by a series of 8 charts each, the hues within each interval 
being obtained by mixtures of the pigments representing the extremes of the in- 
terval. The first chart in a series shows one extreme pigment at the upper right- 
hand corner, the other at the lower left. The upper left corner is white, the lower 
right a mixture in about equal proportions of the two extreme pigments. The 
samples intermediate on the chart represent colors which are intermediate, and 
the color steps between successive samples have been adjusted so that they are, in 
general, about equal. Each chart, therefore, shows a progression of hues from 
the upper right to the lower left corner, while the hues of the samples along the 
lines parallel to the other diagonal are nearly constant. The second chart in a 
series differs from the first by admixture of a gray ink; the third chart corresponds 
to admixture of a darker gray ink; and so on down to the eighth chart which shows 
very dark colors. An alphabetical list of about 4000 color names is given to- 
gether with a key by means of which each corresponding sample may be found in 
the charts. These samples are also identified by name in the charts themselves. 
The large number of the samples makes the steps between successive colors so 
small that interpolation is often not necessary. On this account they may be 
used conveniently as practical color standards in nearly every field. Further- 
more, the scholarship and thoroughness of treatment have given this work a wide 
reputation as the foremost authority on color names." 

For the purposes for which this work was designed and issued, it 
leaves little to be desired. It was intended primarily as a dictionary 

* The more important ones of American origin are briefly described in letter 
circular LC-481 of the National Bureau of Standards. 4 



April, 1937] ANALYSIS AND SPECIFICATION OF COLOR 393 

of color and not as a color system. As a convenient means of general 
color specification, it is subject to the limitation that the samples are 
not removable and would not seem to be readily duplicable in case of 
damage. 

(2) The Munsell Book of Color. 6 The arrangement of colors in 
this system* follows the idea of Helmholtz, that color has three attri- 
butes or dimensions, designated in the Munsell system as hue, value, 
and chroma. In the O. S. A. Colorimetry Committee report 2 these 
attributes are designated as hue, brilliance, and saturation, respec- 
tively. This three-dimensional classification of colors has wide ac- 
ceptance and both sets of terminology are in common use, the former 
among artists and educators, the latter among scientists and tech- 
nologists. The Colorimetry Committee report and the Munsell Book 
of Color should be consulted for further discussion of these terms. 

The following description of the Munsell books is taken from NBS 
letter circular LC-481 : 4 

"Two editions, standard and abridged, each giving approximately 400 different 
color samples. The standard edition consists of charts, one for each of ten differ- 
ent hues, showing colors varying in lightness (Munsell term : value) and satura- 
tion (Munsell term : chroma) ; there are also charts, one for each of eight chromas, 
showing colors varying in hue and value; charts, one for each of six values, show- 
ing colors varying in hue and chroma; and two charts showing altogether 20 hues 
at maximal chroma for each of eight values. The samples are rectangles of matt 
or nearly matt paper, / by 7 /s inches, except for those of the constant-value 
charts which are l / 2 by 6 / 8 inches. The abridged edition consists of 20 constant- 
hue charts made up of Vz by 6 /8-inch samples. In both editions the samples of 
the constant-hue charts are arranged in rows and columns, the samples in any one 
row being equally light, and the samples in any one column being equally satu- 
rated. The colors progress from very light at the top of each chart to very dark 
at the bottom by steps which are visually equal ; and they progress from achro- 
matic colors at the left side of each chart to saturated colors at the right by steps 
which are also visually equal. Each sample is identified by three symbols the 
first indicating hue, the second, value, and the third, chroma. These charts, be- 
cause of the logical arrangement of the samples and the fact that the color differ- 
ences between successive samples are visually equal, have a wide application; 
they are used in color education, in the setting of color tolerances, and as practical 
color standards." 

It is stated in the standard edition that these papers "will stand 
reasonable use, but they should be protected as much as possible from 
light and from gas fumes." They may be obtained in sheets of vari- 

* The Book of Color is a revision and extension of the Atlas of the Munsell 
Color System. 



394 K. S. GIBSON [J. S. M. p. E. 

ous sizes, or in disks for spinning, so that they are adaptable for use in 
colorimeters or in colorimetric research, as well as providing an atlas 
of color. As a method of specification, the system, as exemplified in 
the Book of Color, is limited in that the large number of highly satu- 
rated colors exhibited by dyes and pigments are not included. Also 
the total number of samples is so limited that considerable interpola- 
tion is, in general, necessary to place a color accurately in the system. 
(3) Lovibond Tintometer Glasses, comprising the "Lovibond colour 
scale," sold by The Tintometer, Ltd. 7 The following description of 
these glasses is taken from a Bureau of Standards scientific paper 8 
which in turn quotes from a description of the system published 
in 1912: 

"There are three sets of colored glasses, red, yellow, and blue the red glasses 
absorbing most strongly in the green, the yellow glasses in the blue-violet, and the 
blue glasses in the orange-red. Each set is numbered by the makers in units from 
1.0 to 20.0 with many subdivisions, especially of the lower numbers. Glasses of 
higher grades may also be obtained. 'Each ordinary scale consists of glass slips 
all of one color, but differing in depth, the divisions of difference being equal 
throughout. . . . 

' 'The color units are not only of equal depth throughout each scale, but have 
also a color equivalence in relation to each other. . . . This equivalence of color 
value is accomplished when a normal white light can be gradually absorbed to 
extinction without the development of any color by successive additions of an 
equal number of units of the red, yellow, and blue glass. . . . The starting point 
of the system is the neutral tint unit, which is the quantity of normal white light 
absorbed by one red, one yellow, and one blue glass standard unit combined. 

' 'The intervals between the units, or main divisions of the scales, are the small- 
est differences between which the normal vision can discriminate in the deeper 
shades of glass colors; these are subdivided into tenths as the shades get lighter, 
and ultimately into hundredths in the very light shades; in fact, increasing in 
minuteness as the discriminating power of the eye increases. The actual dimen- 
sions of the unit are arbitrary and of no real importance, so long as the divisions 
are equal both in dimensions and equivalence, and the subdivisions are sufficiently 
minute for the work required.' " 

The red and yellow Lovibond glasses have been widely used in 
America in the color grading of vegetable oils, as well as of mineral oils 
and other products. They are convenient to use and relatively 
cheap, and are probably permanent in their characteristics with care- 
ful usage. When used with various instruments they provide a use- 
ful method of subtractive colorimetry, although limited somewhat in 
range and precision, as noted in Section IV. 

As a means of color specificatibn any system of material samples, 



April, 1937] ANALYSIS AND SPECIFICATION OF COLOR 395 

such as the three just noted, is subject to certain general limitations. 
These are: 

(1) Lack of Permanence and Reproducibility. If the samples com- 
prising the system, which are often called the "standards," are not 
reasonably permanent, they may be unsafe to use, because of the false 
information they will give. At best they can give only a very limited 
service. If permanent, they may be very useful, but in case of loss or 
damage it may be difficult to duplicate them exactly. 

(2) Restriction of Range. This has already been noted. The im- 
portance of this restriction depends upon the problem. The range of 
colors that can be produced by opaque materials is somewhat less than 
by transparent filters such as glasses and solutions. When, however, 
two or more such filters are used in combination, which would be 
necessary if one's set were not to contain a very large number of 
standards, the transmission of the filters, and therefore the resulting 
brightness, would be so reduced as to decrease seriously the precision 
of the values obtained. In general, therefore, it may be said that 
whereas several of the systems of material standards cover a wide 
range of colors they will not satisfactorily cover the highly saturated 
colors shown by certain dyes and pigments and by the spectrum. 

(3) Lack of Uniformity. All systems of material color standards 
exhibit occasional lack of uniformity in passing from one step to 
another. This reduces the precision of results obtained by their use. 
For example, lack of uniformity among the red Lovibond glasses 
proved so troublesome to the American oil chemists that they re- 
quested the National Bureau of Standards to calibrate these glasses 
for them, and more than 3000 such glasses have been given new 
grades 9 during the past few years.* 

The usefulness of any system such as those listed is greatly im- 
proved when the standards can be used in a suitable type of colorime- 
ter. An example of this is the disk colorimeter 11 employing Mun- 
sell papers, which has been used in the color grading of various agri- 
cultural commodities. This provides continuous variation of color 
by optical mixture of the colors of a few selected papers, thus elimi- 
nating any need for interpolation. Furthermore, by spectropho to- 
metric standardization of the papers used, followed by appropriate 
colorimetric computations, the results may be expressed in terms of 
any desired colorimetric system. 

* For Bureau publications on Lovibond glasses consult letter circular LC-398. 10 



396 K. S. GIBSON [j. S. M. p. E. 

(IV) COLORIMETERS 

Instruments designated as colorimeters will be discussed briefly 
under three headings: (1) Colorimeters dependent upon material 
color standards; (2) colorimeters using spectrum primaries; and (3) 
filter photometers. Such a division is rather arbitrary but will serve 
for the purposes of this paper. 

(1) Colorimeters Dependent upon Material Color Standards. 
These may be additive or subtractive. In an additive colorimeter 
the standards or primaries are mixed in such a way that the mixture 
is the sum of the components ; the individual elements add up to give 
the final result. In a subtractive colorimeter the final result is 
obtained by passing light successively through the standards, each 
absorbing in turn a part of the light transmitted by the previous 
one, until the desired color is obtained. 

An example of the former is the disk colorimeter, employing 
Munsell papers, noted above. The personal equation* can be made 
very small in this instrument by the proper choice of disks. The 
most serious drawback to the method would seem to be the un- 
certainty in the colorimetric values of the Munsell papers used as 
standards, either due to lack of initial certainty in the values or to 
change resulting from handling or other usage. 

A second example of an additive colorimeter is that designed by 
Guild. 12 - 13 In this instrument lights of three different colors, 
obtained by transmission through red, green, and blue dyed gelatin 
filters, are optically mixed in the necessary proportions to match the 
sample. The personal equation with this instrument is doubtless 
intermediate between that of instruments like the disk colorimeter, 
where the two halves of the photometric field are more nearly physi- 
cally matched, and those in which there is an extreme physical 
difference, as when spectral primaries are used. The instrument has 
had considerable industrial use in England but little in this country. 
In a recent paper 14 the method of transforming results obtained on 
this instrument to results expressed in terms of the 1931 I.C.I, co- 
ordinate system** is explained in detail. 

An additive trichromatic colorimeter has recently also been de- 



* This expression is used to describe the differences in results obtained by even 
normal observers, because of differences in color vision, when matching colors 
evoked by stimuli of differing physical composition. 
** Vide infra, Section VI. 



April, 1937] ANALYSIS AND SPECIFICATION OF COLOR 397 

scribed by Newhall, 15 who likewise outlines the procedure for convert- 
ing values obtained thereby to the 1931 I.C.I, coordinate system. 
The instrument was designed especially for investigating color per- 
ception, adaptation, after-images, and allied transient phenomena. 
It functions as a combined visual stimulator and trichromatic 
colorimeter. 

The Lovibond glasses when used in a suitable optical instrument 
afford an example of subtractive colorimetry. As they have often been 
used, for example, by the oil chemists in the color grading of oils, there 
has been no adequate means provided for independently varying the 
brightness of the field. The Lovibond scale is itself arbitrary; but, 
as on the Munsell scale, the results can, if desired, be expressed in 
terms of more fundamental colorimetric quantities. The glasses are 
probably permanent with careful usage. The personal equation de- 
pends, as usual, upon the degree of physical match in the two energy 
distributions whose colors are being compared. 

Other means of subtractive colorimetry are available such as the 
use of glass wedges, 16 dyed gelatin wedges, 17 or solutions. 18 

It may be emphasized that the general reliability of such colorime- 
ters, additive or subtractive, is dependent upon (1) the accuracy 
with which the colorimetric values of the material standards can be 
determined and maintained; (2) the reduction in the personal equa- 
tion through similarity of the spectral transmissive or reflective 
properties of the sample and standard; and (3) the correctness of the 
optical and mechanical design of the instrument. 

(2) Colorimeters Using Spectrum Primaries. Such instruments 
have the advantage of eliminating the material color standards that 
form so essential a part of the instruments already described. They 
are usually of two types, the "trichromatic" colorimeter and the 
"monochromatic" colorimeter, both additive. In the first type the 
color of the sample is matched by a mixture of light from three differ- 
ent wavelengths in the spectrum. An example of this is the instru- 
ment designed by Verbeek, 19 in which red light of wavelength 700 m^, 
obtained by means of an incandescent lamp and filter, is optically 
mixed with light of wavelengths 546.1 m/i in the yellow-green and 
435.8 m/x in the blue-violet, obtained from mercury lamps. Tri- 
chromatic colorimeters have also been designed and used by Wright, 20 
and by Guild. 21 

In the so-called "monochromatic' 1 colorimeter, either light from the 
sample is matched by light from a small wavelength band in the 



398 K. S. GIBSON [J. s. M. p. E. 

spectrum mixed with light from some heterogeneous stimulus often 
designated as "white light;" or, if the sample is purple, the spectrum 
light is added to the sample light to match the "white" light. Two 
designs of colorimeters of this type may be cited: (1) that designed by 
Nutting 22 ' 23 and that designed by Priest. 24 

The disadvantages of colorimeters of the type described in this 
section are that (1) the personal equation is at a maximum, and (2) 
the cost and complexity of the design are greater than those of other 
types. In general, these instruments are more suited to research 
than routine testing or control purposes and have probably been 
used most in investigations on the colorimetric properties of the eye. 

(3) Filter Photometers. There are many instruments on the market 
that are called colorimeters, or are designated by some term of which 
"color" is a part, but that should more properly be called compara- 
tors or filter photometers. So-called colorimeters of the Duboscq 
type, widely used in chemical analysis, may be put in this classifica- 
tion. In such instruments the color of one side of a comparator field 
is adjusted, usually by varying the length of optical path through a 
solution, until it is brought to the nearest color match with the other 
side of the field, whose color is produced by a standard solution or fil- 
ter. In some cases filters isolating more or less narrow regions of the 
spectrum are used over the eyepiece of the instrument and a bright- 
ness match used as the criterion. 

Another type of visual filter photometer includes those designed 
primarily for transmission or reflection measurements. Such are the 
Ives tint photometer, 25 employing five filters designated as red, 
yellow, green, blue-green, and blue-violet; the Priest-Lange reflec- 
tometer, 26 ' 27 using the Martens polarization photometer; and the 
Pfund multiple-reflection instruments, 28 in which small color differ- 
ences are accentuated by multiple reflections. 

Several instruments are also on the market of similar kind except 
that a photoelectric cell is the detector instead of the eye. Examples 
of these are the Toussaint photo-colorimetre, 29 employing six filters 
and yielding data for a rough spectrophotometric curve; the General 
Electric "brightness tester" 30 used with blue filter to measure the 
"brightness" of paper; and the Hunter photoelectric reflectometer, 31 
in which an effort is being made to use three filters of such character- 
istics as would justify calling the instrument a colorimeter. Further 
information regarding photoelectric "colorimeters" will be found in 
letter circular LC-473 32 of the National Bureau of Standards, in which 



April, 1937] ANALYSIS AND SPECIFICATION OF COLOR 399 

are discussed the principles underlying the use of photoelectric instru- 
ments as simple photometers, filter photometers, and colorimeters. 
It may be noted that many of the claims made in the advertising 
literature with respect to photoelectric "colorimeters," etc., are mis- 
leading and unjustified. 

(V) SPECTROPHOTOMETRY 

One may well ask at this point whether any method of colorimetry 
is available that will avoid the defects apparent in the various color 
systems and colorimeters exemplified in the previous sections. These 
defects, it will be recalled, are lack of permanence, uniformity, or 
reproducibility of the material color standards or filters used, limita- 
tion of range of colors included by the system or instrument, and the 
personal equation resulting from differences in color vision which may 
cause results by normal observers to differ importantly and by color- 
blind observers to be utterly unreliable. This query leads at once to 
a consideration of the most fundamental method of colorimetry 
available, viz., that by way of spectrophotometric measurement and 
colorimetric computation. 

In the filter photometers noted above certain spectral regions are 
isolated by means of filters. In a spectrophotometer the isolation is 
accomplished with a prismatic dispersing system, by means of which 
a spectrum is formed, and the desired region is selected for use by 
means of entrance and exit slits. A photometric device is added to or 
incorporated with the spectrometric part, and these, combined with 
the proper illuminant and with means for holding the samples, com- 
prise the usual spectrophotometer. Both visual and photoelectric 
methods are in common use in the visible spectrum. 

It is unnecessary to enter into detailed discussion of spectro- 
photometers or spectrophotometric methods in this paper because the 
subject is sufficiently treated in three publications to which reference 
may be made. The first of these is the Optical Society of America 
Spectrophotometry Committee report, 34 published in 1925. This re- 
port deals with the nomenclature of Spectrophotometry and discusses 
in detail the various factors entering into spectrometry and pho- 
tometry, two subjects with which one should be thoroughly familiar 
in order to conduct accurate spectrophotometric measurements. 
Various visual instruments are illustrated and briefly discussed, to- 
gether with various auxiliary methods of Spectrophotometry used to 
supplement the visual method. While slightly out of date in certain 



400 K. S. GIBSON [J. s. M. P. E. 

parts, it still is recommended for those who wish to become familiar 
with the fundamentals of spectrophotometry. 

The other two papers to which reference is made discuss the 
subjects of photoelectric and visual spectrophotometry, respec- 
tively. In the first of these 35 the subject of photoelectric spectro- 
photometry is reviewed from its beginning, near 1900, to 1930. The 
advantages and limitations of amplified-current instruments are dis- 
cussed, in addition to the various types of errors to which all spectro- 
pho tome trie methods, including the photoelectric, are liable. In the 
second paper 36 typical present-day visual instruments and methods 
are described and illustrated and various factors affecting the re- 
liability of spectropho tome trie data are discussed. 

In addition to these general references mention should be made of 
the most recent design of recording photoelectric spectrophotome- 
ter, 37>38 which incorporates two important features not found in 
most spectropho tometers. These are (1) the second prismatic dis- 
persion, which effectively eliminates errors in measurement arising 
from stray energy of other wavelengths, and (2) the dispersion of the 
energy and the selection of the desired wavelength band prior to its 
incidence upon the sample, which avoids undue heating of the 
sample during measurement. 

The principal advantage of a spectrophotometer over a filter 
photometer is the accurate analysis that is possible with the former in 
the measurement of transmission or reflection quantities as a function 
of wavelength. In filter photometry one is always limited by the 
number of available filters; he does not have complete freedom in the 
selection of wavelengths, and the spectral regions transmitted by the 
filters are in most cases relatively broad. With the spectrophotome- 
ter, on the other hand, one can set accurately at any wavelength 
in the spectrum to which the instrument and detector are suited, and 
he can by means of accurate bilateral slits restrict the wavelengths 
transmitted to a very small range, so small in most cases that further 
decrease would not significantly affect the value obtained. Over the 
wavelength range for which the instrument is designed and to which 
the detector is adequately sensitive, therefore, the spectrophotometer 
affords a complete spectral analysis of the radiant energy transmitted 
or reflected by any given object for those angular conditions of illumi- 
nation and viewing used on the instrument. 

We may next consider the purpose of spectrophotometry. This is 
usually one or other of the following: (1) The spectral transmissive 



April, 1937] ANALYSIS AND SPECIFICATION OF COLOR 401 

or reflective data may be the sole purpose of the measurements. The 
data obtained are a fundamental property of the material and afford 
an extremely valuable supplement to other chemical and physical 
analyses. As such, one is often interested in the ultraviolet and 
infrared regions of the spectrum, as well as the visible region, and 
other methods photographic and radiometric are used to extend 
and supplement the visual and photoelectric methods. Since, how- 
ever, this application of spectrophotometry is not under consideration 
in the present paper, it will be given no further attention here. (2) 
The spectral transmissive or reflective data may serve as a funda- 
mental basis from which a minimum of three numbers may be com- 
puted serving to specify the color of the material under certain 
specified conditions and for an assumed average normal observer. 
Such computations will now be considered. 

(VI) COLORIMETRIC COMPUTATIONS 

It may be emphasized that, while a spectral transmission or re- 
flection curve is the most fundamental basis for a colorimetric speci- 
fication, it does not in itself specify the color except in a very crude 
way. It is true that with a little experience one can estimate from 
such a curve what the color is, whether red, yellow, dark green, purple, 
brown, etc., and can also often judge qualitatively as to the difference 
in color corresponding to two curves. But it is quite impossible for 
one to define the color in terms adequate for precise specification ; or 
in similar terms to describe the difference between the two colors 
corresponding to two curves that differ only slightly. 

The trichromatic nature of normal color vision has been noted 
several times above. It is apparent in the hue-brilliance-saturation 
classification of color by the O.S.A. Colorimetry Committee and the 
hue-value-chroma arrangement of the Munsell color system, in the 
three series of colored glasses comprising the Lovibond color system, 
and in the various trichromatic colorimeters. It is further exempli- 
fied in the colorimetric system established by the International Com- 
mission on Illumination in London in 193 1. 39 At that time data 
were adopted defining a hypothetical normal observer and a tri- 
chromatic coordinate system by means of which spectrophotometric 
values may be converted by computation into three numbers serving 
as a specification of the color. The data so adopted define "The 1931 
I.C.I, standard observer and coordinate system." They are de- 



402 K. S. GIBSON [J. s. M. p. E. 

scribed in detail in papers by Smith and Guild 14 and by Judd, 40 
and in the Handbook of Colorimetry. 1 

The methods of converting spectrophotometric data into colori- 
metric terms are so fully described in these publications that it is un- 
necessary to describe them here. The end result of the computations 
is to obtain a minimum of three numbers which will serve to specify 
the color. This specification may be given in any of three com- 
monly used forms : 

(1) Values designated as X, Y, Z may be computed, which are the 
amounts of the I.C.I, primaries (red, green, and blue, respectively) 
required by the standard observer to match the stimulus whose color 
is to be specified. 

(2) If 5 be the sum of X, Y, and Z, the trilinear coordinates (also 
called trichromatic coefficients) are defined as x = X/S, y = Y/S, 
z = Z/S. The values of x, y, and z serve to specify what is termed 
the chromaticity of the color, which is that quality of the color de- 
termined by its hue and saturation, but not by its brilliance or light- 
ness. Since x + y -f z = 1, any two of these numbers serve to specify 
chromaticity. The complete colorimetric specification consists, 
therefore, in giving values of (usually) x and y, in addition to a third 
term, identical with Y above, representing the luminous value of the 
color, such as the luminous transmission or apparent reflectance of the 
sample. 

(3) It is sometimes desirable, in place of x and y, to compute 
what are known as the dominant wavelength, A, and the purity, p, 
which may be derived from the quantities x and y and certain other 
colorimetric data. The dominant wavelength and purity of the 
stimulus correlate loosely with the hue and saturation of the color and 
may thus be more immediately comprehended than the trichromatic 
coefficients. As a means of specification, however, they are somewhat 
less flexible to use than are the values of x and y. When the method 
is used, the complete specification consists of giving values of A, p, 
and F. 

(VII) COLORIMETRIC SPECIFICATIONS 

A complete colorimetric specification suitable for use both in the 
standardizing laboratory and in the factory or control laboratory will 
usually consist of three parts : 

(1) There must be a fundamental definition of the color or range 
of colors that it is desired to specify. This is afforded by giving 



April, 1937] ANALYSIS AND SPECIFICATION OF COLOR 403 

values of x, y, and Y, as above, or of other equivalent values. These 
being expressed in terms of the 1931 I.C.I, standard observer and 
coordinate system, or in terms of any other defined observer and 
system, one can tell by spectrophotometric analysis and colorimetric 
computation whether the sample meets the specification, and there 
need be no important uncertainty in the result except that due to un- 
certainty in the spectrophotometric data. Questions as to per- 
manence of material standards and personal equations of individual 
observers do not enter, and there is no restriction as to range of 
colors. Such a specification is necessary before a standardizing 
laboratory can act as referee in a disputed case. 

However, such a specification is of little practical value to the manu- 
facturer or inspector who must control the quality of his output. At 
present the only general, satisfactory method in such cases is the use 
of working standards, of as great permanence as possible. The ques- 
tion of tolerances then becomes of immediate importance. An ex- 
ample may assist toward a clearer understanding of what is involved. 

Consider the green glasses used in railway signalling. The manu- 
facturer must have a certain range of color within which his product 
will be acceptable, if the cost of manufacture is not to be excessive. 
Furthermore, since only six colors are used in railway signalling there 
is no reason why a considerable range of colors should not be accept- 
able as "green." Obviously, however, there must be a yellow limit 
to avoid confusion with the yellow signal, a blue limit to avoid con- 
fusion with the blue, a pale limit to avoid confusion with white, and a 
dark limit to prevent the signal from being of too low intensity. The 
manufacturer or inspector who is not equipped to test according to the 
type of specification given under (1), above, must therefore have 
"limit glasses," by means of which he may test the quality of his out- 
put. In order that one manufacturer shall not be penalized relatively 
to another it is obviously important that the various limit glasses of 
any one kind going to different manufacturers shall be extremely close 
duplicates. 

(2) A second specification is therefore necessary, setting tolerances 
within which the limit glasses must come. This specification must 
also be expressed in terms of x, y, and Y, or similar quantities. 

(3) A third specification is then necessary, outlining the procedure 
to be followed by the manufacturer or inspector on testing the output 
by means of the limit glasses. 

An example of this three-part type of specification is found in 



404 K. S. GIBSON [J. S. M. P. E. 

A.A.R. Signal Section Specification 69-35.* 1 It is of course true that 
a specification as elaborate as this need not in all cases be formulated. 
In some cases merely the fundamental definition with tolerances may 
be sufficient. It should be noted, however, that a specification with- 
out tolerances, while adequate in most cases as a record of a color, will 
usually not prove of practical value for testing purposes. 

(VIII) COLOR OF ILLUMINANTS 

So far we have considered only the colorimetry of illuminated ob- 
jects and have for the most part passed over the colorimetry of 
illuminants. This, however, has been a matter of convenience in 
presentation rather than indicating a lack of importance. Indeed, the 
most accurate colorimetry of objects requires the use of a standardized 
illuminant. Natural daylight is greatly used, of course, in the direct 
comparison of the colors of two objects, this including the use of the 
various color systems mentioned in Section III, but this illuminant is 
often unsatisfactory because of variability in both chromaticity and 
intensity. Artificial daylight is therefore extensively used for this 
purpose. 

Likewise the most accurate use of any of the colorimeters noted in 
Section IV requires the use of a standard illuminant, both to illumi- 
nate the sample and to illuminate the standards or calibrate the 
primaries. An illuminant standardized with respect to color or 
spectral energy distribution is not required for accurate spectro- 
photometry, but is required for the colorimetric computations based 
upon the spectrophotometric data. 

For such computations, as well as for use in direct colorimetry, the 
International Commission on Illumination, at the same time that 
data were adopted defining the standard observer and coordinate 
system in 1931, also defined and recommended the use of three stand- 
ard illuminants, designated as I.C.I, illuminants, A, B, and C, re- 
spectively. 

Illuminant A is defined as a gas-filled lamp operated at a color- 
temperature of 2848K. The term "color-temperature" is defined 
below. 

Illuminant B is defined as the same lamp (i. e., illuminant A) used 
in combination with a specified Davis-Gibson filter, the light trans- 
mitted by the filter having a color and spectral energy distribution 
approximating those of both 4800 K and average noon sunlight 
(skylight excluded). 



April, 1937] ANALYSIS AND SPECIFICATION OF COLOR 405 

Illuminant C is defined as illuminant A combined with a second 
similar filter, the transmitted light in this case having a color and 
energy distribution approximating those of both 6500 K and average 
daylight as given by an overcast sky or by the combined sunlight and 
skylight on a horizontal surface. 

The filters specified for use in I.C.I, illuminants B and C are two of 
an extensive series of reproducible liquid filters designed 42 ' 43 for use in 
colorimetry and photographic sensitometry, one of them having been 
also adopted by the 8th International Congress of Photography for 
use with an illuminant at 2360 K to give standard artificial sunlight 
for use in photographic sensitometry. 

Color- temperature is defined as the temperature* at which the ideal 
Planckian radiator or "black body" must be operated to produce the 
chromaticity of the color in question. The importance of this con- 
cept is that the chromaticities of most illuminants can be specified in 
terms of a single variable, accurately in the case of incandescent 
illuminants and approximately in the case of natural and artificial 
sunlight and daylight. One can readily visualize this color- tempera- 
ture scale by remembering the color-temperatures of certain well- 
known illuminants the candle and kerosene flames near 2000 K, 
the acetylene flame and vacuum tungsten lamps near 2400 K, gas- 
filled lamps from 2700 to 3200K, the bare carbon arc near 3700K, 
average noon sunlight and daylight near 5000 and 6500 K, respec- 
tively, and blue sky from 10,000 to 25,000K. 

An illuminant such as the mercury arc can not be specified in terms 
of color- temperature, as it departs too much from the Planckian locus. 
However, sources such as sunlight and daylight, which have colors 
close to but usually not exactly on the Planckian locus can be specified 
on the color-temperature scale by computational methods derived by 
Davis 45 and Judd. 46 

Experimentally, color-temperature may be determined by direct 
comparison with lamps standardized to give color-temperatures in 



* The spectral energy distribution given by Planck's equation for any specified 
temperature depends upon the value of the constant C 2 . The color temperatures 
of 2848 and 2360 K, stated above in connection with I.C.I, illuminants A, B, 
and C, and with the photographic standard of intensity, were defined with Ct = 
14350. On the basis of the new color-temperature scale recently established at 
the National Bureau of Standards 44 with C 2 = 14320, in agreement with the in- 
ternational temperature scale, the same respective energy distributions will be 
obtained at temperatures of 2842 and 2355 K. 



406 K. S. GIBSON [J. S. M. P. E. 

terms of voltage or current; or, above the range where a standard 
lamp can be safely operated, by comparison with standard lamp and 
filter combinations. Such filters may be those of the Davis-Gibson 
series, just noted, carefully selected daylight glass or dyed-gelatin 
filters, or the quartz-nicol combinations used on the rotatory-disper- 
sion colorimetric photometer. 47 

(IX) CONCLUSION 

In concluding this brief review of colorimetric methods we may 
recall that increasing refinement of colorimetric specification is 
afforded by (1) color names, (2) color systems, (3) colorimeters, and 
(4) spectrophotometric analysis followed by colorimetric computa- 
tion. The last method is undoubtedly the only fundamental one; 
the first three partake of a fundamental nature only to the extent 
that they are based upon the spectrophotometric method in any 
particular case. This is not to say that methods 1 to 3 should not 
be used when desirable ; it means rather that, when used, they should 
be based upon method 4 to an extent that in reality makes them 
largely equivalent to method 4. For example, consider the system of 
color names previously referred to,* which is being used to name the 
colors of drugs and chemicals. Each of these names represents a 
certain "pocket" in the "color solid," the boundaries of which will be 
specified in terms convertible to values of x, y, and Y for the 1931 
I.C.I, standard observer and coordinate system. 

Likewise, as a result of spectral apparent reflectance measurements 
on a large number of Munsell samples 48 followed by the usual colori- 
metric computations, it is now possible to convert a colorimetric 
specification in the Munsell system to a specification in terms of this 
same 1931 I.C.I, standard observer and coordinate system. Due to 
slight imperfections and lack of uniformity in the Munsell system such 
a conversion can not at present be made with the greatest precision. 
However, a movement is under way to "idealize" this system by 
eliminating the obvious irregularities now existing therein. If this 
were done, colorimetric specification could be made with equal sound- 
ness upon either the Munsell or the I.C.I, basis, for they would be 
interconvertible, both based in reality upon spectrophotometric 
measurements and computations in terms of the I.C.I, standard 
observer. 

* Section II. 



April, 1937] ANALYSIS AND SPECIFICATION OF COLOR 407 

Examples of colorimeters that may be used to obtain results upon a 
fairly sound basis by means of proper calibration were illustrated by 
the disk colorimeter and the Guild trichromatic colorimeter. 

A further word may be said about the 1931 I.C.I, standard observer 
and coordinate system. There are two outstanding advantages of 
this system. First, the chromatic and luminous characteristics of the 
average normal observer are so accurately incorporated in this 
hypothetical observer that revision will probably be unnecessary for 
a long time. Second, the international standing of this system en- 
ables people in every country to speak a common language of color 
specification. 

However, the I.C.I, coordinate system, not the standard observer, 
has one serious defect that should not be overlooked. Distances on 
the (x,y) diagram do not properly represent chromaticity differences 
as seen by the normal eye. This has been pointed out by Judd, 49 
who has proposed an equal-chromaticity-scale coordinate system 
which is a transformation of the I.C.I, system. While it should be 
subjected to further practical tests and will probably need some 
revision, it is undoubtedly a considerable improvement over the 
I.C.I, coordinate system in its correlation with visual judgments. 
This is of great importance in the specification of tolerances. 

In deciding upon the method to use for the measurement and speci- 
fication of color stimuli in any particular problem one must consider 
the two elements of time and accuracy. Lack of time has often been 
the excuse for the employment of loose methods of colorimetry and 
color specification. The time required for accurate spectrophoto- 
metric measurement and colorimetric computation has often made 
such a method impossible to use. And indeed there are many colori- 
metric problems in the control laboratory that do not need any 
fundamental method of measurement and specification. However, 
the commercial development of recording photoelectric spectro- 
photometers may so reduce the time of measurement that this 
method may become a strong competitor with less exact methods 
even in the control laboratory. 

In closing, one important defect (for many purposes) of practically 
all spectrophotometers may be noted. This defect is that but 
one set of angular conditions of illuminating and viewing are available 
on any one instrument now commercially available. A consideration 
of these angular conditions is of particular importance in the analysis 
and specification of the colors of glossy materials. The colors of such 



408 K. S. GIBSON [j. s. M. P. E. 

materials, for example, a glossy red enamel, change notably under 
differing conditions of examination. Diffuse illumination may mask 
differences between two samples readily apparent by ordinary inspec- 
tion when one can examine the samples under various angles of il- 
lumination and viewing. No answer to this problem is attempted 
in this paper. Gloss specification is in itself of great complexity. 50 
The 45-normal conditions of illuminating and viewing, recom- 
mended by the I.C.I., 14 ' 39 are probably the best single set of condi- 
tions to use for glossy materials, but complete colorimetric speci- 
fication of glossy chromatic materials has not yet been accomplished. 
The subject is too important to be ignored, however. Until it is 
solved, direct visual comparison with permanent working standards 
must be used to some extent for such materials. 

Rapid progress will doubtless be made along this line, as is being 
made along other lines. The interest in color measurement and 
specification is much greater than ever before. An example of this 
interest is afforded in the hundreds of requests that were received at 
the Bureau for a recent letter circular 32 on photoelectric colorimeters 
and by the fact that several hundred copies of the letter circular 4 on 
color charts have been issued to applicants during the past three 
years. 

Those in the motion picture industry have their own color problems 
to solve. The material in this paper may not at first seem to have 
much practical application in the development of accurate color 
reproduction in motion pictures. Those in charge of such develop- 
ment can not fail, however, to benefit by a knowledge of the funda- 
mentals of colorimetry to which reference is made in this paper. 

REFERENCES 

1 The Technology Press, Massachusetts Institute of Technology, Cambridge, 
Mass. (1936). 

2 Report of the Colorimetry Committee, L. T. Troland, Chairman, J. Opt. 
Soc. Amer. and Rev. Sci. Instr., 6 (Aug., 1922), No. 6, p. 527. 

3 LUCKIESH, M.: "Color and Its Applications," D. Van Nostrand Co., 
New York (1915). 

4 "Color Charts," Letter Circular LC-481, National Bureau of Standards, 
Washington, D. C. 

5 MAERZ, A., AND PAUL, M. R.: "Dictionary of Color," McGraw-Hill Book 
Co., New York, N. Y. (1930). 

1 "Munsell Book of Color," Munsell Color Co., Baltimore, Md. (1929). 

' "Colour Measurement," The Tintometer, Ltd., Salisbury, England. 

8 GIBSON, K. S., AND HARRIS, F. K.: "The Lovibond Color System; I. A 



April, 1937] ANALYSIS AND SPECIFICATION OF COLOR 409 

Spectrophotometric Analysis of the Lovibond Glasses," Scientific Paper No. 547, 
National Bureau of Standards, Washington, D. C. 

HAUPT.. G. W.: "Statistical Investigation of the Uniformity of Grades of 
Lovibond Red Glasses," /. Opt. Soc. Amer., 27 (Jan., 1937), No. 1, p. 63. 

10 "Publications on Colorimetry and Spectrophotometry," Letter Circular 
398, National Bureau of Standards, Washington, D. C. 

11 NICKERSON, D.: "A Colorimeter for Use with Disk Mixture, and Color 
Measurements in Psychological Terms," J. Opt. Soc. Amer., 21 (Oct., 1931), No. 
10, p. 640. 

11 GUILD, G.: "A Trichromatic Colorimeter Suitable for Standardization 
Work," Trans. Opt. Soc. London, 27 (1925), No. 2, p. 106. 

18 "The Guild Colorimeter," A dam Hilger, Ltd. 

14 SMITH, T., AND GUILD, J.: "The C.I.E. Colorimetric Standards and Their 
Use," Trans. Opt. Soc. London, 33 (1931), No. 3, p. 73. 

16 NEWHALL, S. M.: "An Instrument for Color Stimulation and Measure- 
ment," Psychological Monographs, 47 (1936), No. 212, p. 199. 

16 JUDD, D. B.: "A Subtractive Colorimeter for the Measurement of Small 
Chromaticity Differences between Surfaces of Moderate Spectral Selectivity of 
Reflectance," /. Opt. Soc. Amer., 26 (May, 1936), No. 5, p. 225. 

17 JONES, L. A.: "A: Colorimeter Operating on the Subtractive Principle," 
/. Opt. Soc. Amer., 4 (Nov., 1920), No. 6, p. 420. 

18 ARNY, H. V., AND TAUB, A. : "Standardized Colored Fluids and Some Offi- 
cial Colorimetric Tests," J. Amer. Pharm. Assoc., 12 (Oct., 1923), No. 10, p. 839. 

19 VERBEEK, H. P. J.: "Een Trichromatische Colorimeter," Physica, 13 
(1934), No. 3, p. 77. 

80 WRIGHT, W. D.: "A Trichromatic Colorimeter with Spectral Primaries," 
Trans. Opt. Soc. London, 29 (1927), No. 5, p. 225. 

21 GUILD, J.: "The Colorimetric Properties of the Spectrum," Phil. Trans. Roy. 
Soc. (London), A230 (1931), p. 149. 

"NUTTING, P. G.: "A New Precision Colorimeter," Bull. Bureau of Stand- 
ards, 9, 1 (1913); B. S. Sci. Paper No. 187. 

11 "The Nutting Colorimeter," A dam Hilger, Ltd. 

24 PRIEST, I. G. : "Apparatus for the Determination of Color in Terms of Domi- 
nant Wavelength, Purity, and Brightness," J. Opt. Soc. Amer., and Rev. Sci. 
Instr., 8 (Jan., 1924), No. 1, p. 173. 

25 Descriptive material issued by the Palo Co., New York, N. Y. 

2 * PRIEST, I. G.: "The Priest-Lange Reflectometer Applied to Nearly White 
Porcelain Enamels," J. Research, National Bureau of Standards, 15 (Nov., 1935), 
No. 5, p. 529. 

27 Descriptive circular issued by Akatos, Inc., New York, N. Y. 

28 Descriptive material issued by the Munsell Color Co., Baltimore, Md. 

29 Descriptive material published by J. Carpentier, Paris. 

30 DAVIS, M. N.: "Instrumentation in Brightness Grading," Paper Trade J., 
101 (July 4, 1935), No. 1, TS 36. 

LEWIS, L. C.: "Definition of Brightness," Paper Trade J., 101 (Aug. 8, 
1935), No. 6, TS 64. 

31 Descriptive material issued by H. A. Gardner Laboratory, Washington, 
D. C. 



410 K. S. GIBSON 

32 "Photoelectric Colorimeters," Letter Circular LC-473, National Bureau of 
Standards, Washington, D. C. 

33 GIBSON, K. S. : "Photoelectric Photometers and Colorimeters," Instruments, 
g (Nov., Dec., 1936), No. 11, 12, pp. 309, 335. 

34 "Spectrophotometry," Report of O. S. A. Progess Committee for 1922-23 
(K. S. Gibson, Chairman), J. Opt. Soc. Amer. and Rev. Sci. Instr., 10 (Feb., 1925), 
No. 2, p. 169. 

35 GIBSON, K. S.: "The Use of the Photoelectric Cell in Spectrophotometry, 
Photoelectric Cells and Their Applications," Phys. and Opt. Societies (London). 
Also obtainable from Adam Hilger, Ltd., London (1930). 

36 GIBSON, K. S.: "Visual Spectrophotometry," J. Opt. Soc. Amer., 24 (Sept., 
1934), No. 9, p. 234. 

37 HARDY, A. C.: "A New Recording Spectrophotometer," /. Opt. Soc. Amer., 
25 (Sept., 1935), No. 9, p. 305. 

38 MICHAELSON, J. L., AND LiEBHAFSKY, H. A.: "A New Spectrophotometer 
and Some of Its Applications," Gen. Elect. Rev., 39 (Sept., 1936), No. 9, p. 445. 

39 Proceedings, International Commission on Illumination (1931), p. 19. 

40 JUDD, D. B.: "The 1931 Standard Observer and Coordinate System for 
Colorimetry," /. Opt. Soc. Amer., 23 (Oct., 1933), No. 10, p. 359. 

41 A.A.R. Signal Section Specification 69-35, Assoc. Amer. Railroads, New 
York, N. Y. 

42 DAVIS, R., AND GIBSON, K. S.: "Filters for the Reproduction of Sunlight 
and Daylight and the Determination of Color-Temperature," Miscellaneous Pub- 
lication No. 114, National Bureau of Standards (1931). 

43 DAVIS, R., AND GIBSON, K. S.: "Artificial Sunlight for Photographic Sen- 
sitometry," Trans. Soc. Mot. Pict. Eng., XII (1928), No. 33, p. 225. 

44 WENSEL, H. T., JUDD, D. B., AND ROESER, W. F.: "Establishment of a 
Scale of Color -Temperature," Bureau of Standards J. of Research, 12 (May, 1934), 
No. 5, p. 527; RP677. 

45 DAVIS, R.: "A Correlated Color-Temperature for Illuminants," Bureau of 
Standards J. of Research, 7 (Oct., 1931), No. 4, p. 659; RP365. 

46 JUDD, D. B.: "Estimation of Chromaticity Differences and Nearest Color- 
Temperature on the Standard 1931 I.C.I. Colorimetric Coordinate System," J. 
Opt. Soc. Amer., 26 (Nov., 1936), No. 11, p. 421. 

47 PRIEST, I. G.: "The Colorimetry and Photometry of Daylight and Incan- 
descent Illuminants by the Method of Rotatory Dispersion," /. Opt. Soc. Amer. 
and Rev. Sci. Instr., 7 (Dec., 1923), No. 12, p. 1175. 

48 GLENN, J. J., AND KILLIAN, J. T.: "Trichromatic Analysis of the Munsell 
System" (Unpublished thesis), Mass. Inst. of Tech., Cambridge, Mass. (1935). 

49 JUDD, D. B. : "A Maxwell Triangle Yielding Uniform Chromaticity Scales," 
National Bureau of Standards, J. of Research, 14 (Jan., 1935), No. 1, p. 41; RP- 
756. Also /. Opt. Soc. Amer., 25 (Jan., 1935), No. 1, p. 24. 

60 HUNTER, R. S.: "Methods of Determining Glass," National Bureau of 
Standards J. of Research, 18 (Jan., 1937), No. 1, p. 19; RP958. 



DIRECT RECORDING AND REPRODUCING MATERIALS 
FOR DISK RECORDING* 



A. C. KELLER** 

Summary. Recently materials for direct disk recording and reproducing work 
have been improved so that they are now suitable for many uses. These materials, 
as they are available on the market, are classified chemically into five groups, and 
measurements are given of frequency characteristic, surface noise, life, distortion, 
etc., and data have been taken with both lateral and vertical recording. 

There has been a need for some time for a suitable combined 
recording and record material of the disk type for use in mechanical 
recording. Materials for this work must be suitable for recording, 
and after recording must be capable of being played back satis- 
factorily. As recording materials, they should be capable of good- 
quality recording with relatively simple apparatus. As record 
materials, they must render reproduced sound of good quality and 
low surface noise with simple apparatus, and must also have adequate 
life and be capable of being satisfactorily reproduced with a sturdy 
reproducer. These requirements are, of course, all relative, and 
depend upon the technical character of the work to be done and 
upon the the standards of excellence required. 

Of course, wax cylinders have been, and are being, used for similar 
types of work for which disks of the kind to be discussed in this paper 
can be used, and these are capable of producing good results. Wax 
disks have also been used, but their applications have, in recent 
years, been largely confined to sound-picture studio and other more 
elaborate uses. Even though these cylinders and disks of wax have 
done many types of work satisfactorily, there has been a real need 
for a thin, sturdy, and easily stored disk of a material upon which 
sound can be recorded and reproduced with rugged instruments. 
This statement seems justified by the growing number of suppliers 

* Published by courtesy of, and simultaneously with, the Acoustical Society of 
America. 

** Bell Telephone Laboratories, New York, N. Y. 

411 



412 A. C. KELLER [J. S. M. P. E. 

of a newer type of record, which usually takes the form of a thin disk 
from 10 to perhaps 60 mils thick and of a variety of materials. 

This paper presents data on some of the materials of this type as 
they are now available on the market for general use. As there are 
many possible combinations of recording instruments, reproducing 
instruments, styli, etc., the data will give a rather general idea of 
what can be expected from the various recording materials. The 
materials will be referred to as of certain types of distinguishing 
physical or chemical characteristics, and not by their trade names. 
While it might be interesting to mention the disks by name, the more 
general classification entirely satisfies the needs of this paper. 

The disk record materials that have come to our attention for this 
type of recording may be classified according to the following groups, 
listed in Table I : 

(1) Materials containing cellulose esters. These materials have been mainly 
plasticized cellulose nitrate. All the disks that we have tested that were 
stated to be cellulose acetate or "acetate" by the suppliers, contained a pre- 
dominant amount of cellulose nitrate. 

(2) Metal disks such as aluminum, zinc, lead, pewter, etc. 

(3) Plasticized thermo-setting phenolic resins which are used soft and un- 
cured as recording materials. A baking operation after recording transforms the 
material into a hard record material. 

(4) Various types of resinous compounds other than the cellulosic and pheno- 
lic compositions mentioned above. 

(5) Gelatin and gelatin compositions. 

TABLE I 

Classification of Disk Record Materials 

Group Material 

1 Cellulose esters 

Metal disks (aluminum, zinc, lead, pewter, etc.} 
Plasticized thermo-setting phenolic resins 

4 Resins other than those of group 1 or 3 

5 Gelatin and gelatin compositions 

Except for the metal disks under group 2, some of the materials of 
all these groups are or have been available as coatings on|metal 
disks. A typical coated metal disk will consist of a Vsz-mch alumi- 
num disk with a 5- to 10-mil coating upon either side. Some disks 
are available only as thin sheets of the recording material, and others 
are available also as coatings. A few have also been supplied as 



April, 1937J MATERIALS FOR DlSK RECORDING 413 

coatings upon paper, and this appears to be a somewhat less expen- 
sive method of preparing the disks; but judged by the relatively few 
that are or have been on the market, it is apparently not as easy to 
produce these of equivalent quality of reproduction and ease of use. 

These direct recording and reproducing disks can be recorded in 
a variety of ways. Many of them are recorded with a cutting stylus 
in much the way that waxes are cut in the phonograph industry. 
The fact that the materials are somewhat harder than wax makes the 
use of an advance ball unnecessary. It will be recalled that the 
usual advance ball used in phonograph work is a carefully ground 
portion of a jewelled sphere located adjacent to the cutting stylus of 
the recording instrument and adjustable with respect to it in such a 
way that a predetermined width of groove may be obtained even in 
very soft materials such as wax, etc. The recording styli used for 
direct recording work are of steel, sapphire, or diamond. The life 
of these is a function of the character of the material itself, the clean- 
liness of the material, and the skill of the operators using them. In 
general, sapphire and diamond styli are more costly, but are harder 
and have a longer Me on many materials than the steel styli now 
available. The jewelled styli are, however, more brittle, and for this 
reason are subject to breakage when used with materials that con- 
tain small particles of foreign matter. 

The shavings that occur in the cutting operation are not generally 
removed by a suction tube, as in wax recording. The continuous 
thread of material that is obtained under properly adjusted recording 
conditions is handled in a number of ways. It has been collected, 
for example, by a radial brush resting lightly upon the record surface, 
and it has also been collected when recording by permitting the shav- 
ing to coil up inside the position of the cutting stylus upon the record. 

Some of the coated disks are pre-grooved, in a molding operation, 
by the use of suitable matrices. These disks require somewhat less 
expensive recording equipment in that the usual lead-screw and 
associated mechanism are not required to feed the recording device 
across the surface of the disk during recording. Pre-grooved disks 
have mainly been used for so-called home recording and similar uses, 
whereas many of the other disks discussed here have been used for 
more critical work. All the materials discussed in this paper are 
capable of being reproduced with rugged instruments guided across 
the record surface during reproduction by the recorded grooves 
themselves, without the aid of a lead-screw. The pre-grooved records 



414 A. C. KELLER [J. S. M. P. E. 

are recorded as an embossing operation in which a blunt tool is used 
in contrast to the cutting tool mentioned earlier. The embossing or 
rubbing tool or stylus in the recording operation embosses a somewhat 
wider and deeper modulated groove than the original molded unmodu- 
lated guiding groove. 

The metal disks under group 2 are also available as pre-grooved or 
plain disks. The pre-grooved variety of disks is recorded in the same 
way as just described for the pre-grooved coated disks. The plain 
metal surface disks, however, are not generally cut as were the coated 
disks, but rather embossed with a blunt stylus in the manner used for 
pre-grooved disks. A surface lubricant is often used during the 
recording of metal disks. 

The gelatin disks under group 5 are generally recorded with a 
cutting tool in the manner of disks of groups 1, 3, and 4. After 
recording, gelatin disks are frequently treated chemically in order to 
harden the record surface. These treated surfaces have a somewhat 
longer life in reproduction. 

For reproducing from these direct recording materials, a number 
of stylus shapes and materials have been used. Stylus materials 
have been steel, sapphire, diamond, and various forms of fiber, thorn, 
etc. In general, the type of needle used will depend upon the design 
of the reproducing instrument and the number of times it is wished 
to reproduce the record. It is possible to design reproducing instru- 
ments of small enough mechanical impedance so that pressures of the 
stylus upon the record material can be reduced to a value such that 
any of the stylus materials mentioned are capable of reasonably 
satisfactory reproduction provided that the stylus has a slightly 
rounded point, that is, has a tip radius of the order of 2 mils. Instru- 
ments that may be satisfactorily used at these moderate record pres- 
sures are a refinement over those available for use in the ordinary 
electrical home phonograph. 

Rapid wear of either the stylus or the record, or both, has been 
encountered at the higher record pressures with some combinations 
of record and reproducer stylus materials. This can also be due to 
the higher coefficient of friction between these material combinations. 
For this reason, even with a properly contoured needle (or one that 
has been approximately contoured by playing a few grooves of a 
common phonograph record, which is somewhat abrasive) certain 
combinations of stylus and record materials will require lower pres- 
sures in order not to injure seriously either the stylus or the record 



April, 1937] MATERIALS FOR DlSK RECORDING 415 

in a few play ings. For example, the coefficient of friction between 
a steel stylus and an aluminum record is about 0.3, compared with 
the unusually low value of somewhat less than 0.02 for a fiber stylus 
and an aluminum record. 

Apart from difficulties arising from the coefficient of friction between 
materials, the contour of the stylus is very important. For example, 
a new steel phonograph needle is relatively sharp, and the area of 
contact between it and the record is of the order of a few millionths 
of a square inch; so that with a reproducer resting upon a record, the 
needle point pressures are of the order of 20 to 100 tons per square- 
inch. Pressures of this order will rapidly wear either the stylus, 
the record, or, most probably, both. An abrasive material is included 
in commercial phonograph records in order to grind the needle to 
fit the groove of the record and to prolong the record life. In a 
matter of seconds the needle point pressures of a steel needle playing 
a common phonograph record are reduced to values of the order of 
5 tons per square-inch. It is interesting to note that even these 
lower pressures are very large, and somewhat higher than pressures 
known to be the upper limits for which lubricants will remain between 
surfaces. This is the reason that lubricating phonograph records 
by various substances has rarely shown any effect upon either surface 
noise or record life. Only for instruments that are capable of being 
satisfactorily used at very low record pressures can the effect of 
lubrication be measured. 

The nature of the reproducer stylus material and the stylus design 
will also affect the quality of reproduction even if used at needle 
point pressures for which the effect upon either the record material 
or the stylus is negligible. For example, a needle with a large effec- 
tive mass will tend to reduce the high-frequency response of the 
reproducer, and a similar effect will be obtained by the use of a thin 
or relatively flexible needle. While the amount of reduction of 
high-frequency response will depend upon the effective mass and 
stiffness of the stylus, it will depend also upon the relationship of 
these to the vibratory constants of the reproducing instrument itself 
and upon the relationship of the mechanical impedance of the com- 
plete reproducing instrument to the mechanical impedance of the 
record material. A reduction of high-frequency response is, in 
general, not desirable, because it will tend to make the reproduction 
less natural, at least for many types of recording. However, a 
reduction in high-frequency response will also have the effect of 



416 A. C. KELLER [j. s. M. P. E. 

reducing surface noise, and this may be important under some 
conditions. 

To give a complete picture of record materials of the type being 
discussed and the reproduction from them with all available repro- 
ducers would make a subject beyond the scope of this paper. Typical 
data will rather be shown, taken with some of the materials and instru- 
ments that are at present available. 

Quantitative measurements taken with both lateral and vertical 
recorders and reproducers at linear speeds of 14 to 20 inches per 
second will include: 

(a) Response vs. frequency characteristics. 

(6) Surface noise. 

(c) Wear (life). 

(d) Distortion measurements. 

Frequency Characteristics. The frequency response has been 
measured by recording single frequencies and then measuring the 
reproducer response as the record is played back. The curves that 
follow show the relative response compared with that obtained from 
pressings made from processed matrices. The waxes used for produc- 
ing the comparison matrices were recorded in the same way and with 
the same recorders used for each of the materials under test. Figs. 1, 
2, and 3 are frequency response measurements taken with lateral 
recording. Fig. 4 shows similar measurements with vertical record- 
ing. 

Fig. 1 shows lateral cut frequency characteristics of materials of 
the cellulosic type, which has been called group 1. The recording 
has been done with a steel recording stylus. In general, the response 
differences of this and other groups occur as the frequency is increased. 
Fig. 1 shows the effect obtained with two reproducers of different 
design and vibratory characteristics. Curve Al taken with repro- 
ducer A y a commercial reproducer fitted with a contoured (about 
2-mil tip radius) steel stylus and material No. 1, shows poorer high- 
frequency response than does curve Bl taken with laboratory repro- 
ducer B fitted with a contoured diamond stylus on the same record. 
Material No. 1 is the best of those tested with lateral recording and 
classified under group 1. Fig. 1 shows also the variation of materials 
of this group as supplied by two manufacturers. Curve B2 has been 
taken on another material of the cellulosic group which showed the 
poorest frequency response of the materials tested of this group. A 



April, 1937] 



MATERIALS FOR DISK RECORDING 



417 



considerable difference in response will be noted in the region of 
5000 cps. between these materials, and the better material is down 
very little in this frequency range as compared with the molded 
record. 

Fig. 2 shows lateral frequency response characteristics taken with 
embossed grooves. These curves are relative to measurements 
taken with molded records made from matrices which in turn have 
been made from lateral cut waxes. Curve A 10 has been taken with a 
pre-grooved material of group 4. Prior to recording, the grooves 
were unmodulated and relatively narrow and shallow. After record- 
ing, the grooves were similar to those of a standard phonograph 
record, that is, about 2 x /2 to 3 mils deep and 6 to 7 mils wide. The 
frequency response of curve A 10 shows a relative drop of more than 
30 db. at 4000 cps., much poorer than any other material tested. 




trz 



100 



500 1000 5000 

FREQUENCY IN CYCLES PER SECOND 
FIG. 1. Lateral recording group 1. 



10,000 



Curves B3 and B4 of Fig. 2 show the results with an aluminum 
disk and lateral embossed recordings. This material is probably the 
most widely used at present of the solid metal group. Curve B3 has 
been taken with a plain aluminum disk and curve B4 with a pre- 
grooved aluminum disk. In recording, the aluminum surface is 
often lubricated with a wax such as paraffin, or with kerosene. The 
use of the lubricant produces a record that is several decibels quieter 
than one recorded without a lubricant. These curves were taken 
with the laboratory reproducer B, which is fitted with a diamond 
stylus, and show that with an adequate recording instrument fre- 
quencies up to 4000 to 5000 cps. can be satisfactorily recorded on 
aluminum. The measurements at 5000 cps. and higher were con- 
sidered doubtful because of the high surface noise levels in this 
region. 

In order to show the effect of the reproducer stylus upon the fre- 



418 



A. C. KELLER 



[J. S. M. P. E. 



quency response, curves A3 and A '3 are shown on Fig. 2. Repro- 
ducer A , although somewhat inferior to reproducer B in response and 
mechanical impedance, lends itself to rapid stylus changes and has 
been used for this reason. Curve A 3 is a remeasurement of the record 
used for getting curve B3 using a contoured steel stylus in the repro- 
ducer. Curve A '3 is another measurement of the same record and 
reproducer A fitted with a fiber needle. The difference in response 
in the region of 5000 cps. can be seen to be considerable. Fiber or 




100 



500 



1000 



5000 10,000 



FREQUENCY IN CYCLES PER SECOND 
FIG. 2. Lateral embossed recording. 

thorn needles wear quite rapidly on aluminum records because of 
their mechanical properties. The characteristic shown is about the 
best that can be obtained with needles of this kind and instruments 
now available. As the needle wears, the response decreases at the 
higher frequencies and the distortion products increase rapidly. 
Fiber or thorn needles are frequently badly worn before completely 
playing a single record with a playing time of the order of a few 
minutes. 

Curve B5 of Fig. 3 shows the relative lateral cut recording frequency 
characteristic of a plasticized thermo-setting phenolic material of 
group 3. The response has been taken after baking the record for 



April, 1937] 



MATERIALS FOR DISK RECORDING 



419 



two hours in a small oven at about 360F. During the baking 
operation, there is some flow of the material which undoubtedly 
changes the frequency characteristic of the recording. Curve B5 
was taken after baking. Difficulty was encountered at some of the 
higher frequencies in measuring this material due to high noise levels 
on some of the records. 

Curve B6 of Fig. 3 shows the relative lateral cut frequency response 
characteristic of a resinous material of group 4. This material is 




i 
-L 

O 



''i 
OJ ro 

00 





































































- 


























GROUP 4 






^ 


N 












































\ 


B6 



u 
-10 
-20 




























. 




1 ^ 


































































GROUP 5 








N 


\ 












































) 


B7 



50 100 500 1000 5000 10,000 

FREQUENCY IN CYCLES PER SECOND 
FIG. 3. Lateral recording. 

the best of those examined of this group. This group, however, 
covers a wide range of materials that have very little in common, so 
that curve B6 must be taken as a measurement of one material only 
and not necessarily as illustrative of this group. 

Curve B7 of Fig. 3 shows the relative lateral cut frequency response 
characteristic of the best gelatin compound of group 5 that was 
measured. This group also gave trouble with high noise levels 
interfering with the measurements at higher frequencies. 

Fig. 4 shows the relative frequency response measurements taken 
with materials of groups 1, 3, 4, and 5 with vertical cut recording. 
As with the measurements with lateral recording, these data have 



420 



A. C. KELLER 



[J. S. M. P. E. 



been plotted to show the relative frequency response compared with 
that obtained from molded records prepared from processed matrices. 
Here again, the waxes used for producing the matrices were recorded 
in the same way and with the same recorders used for each of the 
materials under test. Measurements are not shown for materials of 
group 2 with vertical recording. Although only a limited amount 




100 500 1000 5000 

FREQUENCY IN CYCLES PER SECOND 
FIG. 4. Vertical recording. 



10,000 



of work has been done recently with embossing vertically modulated 
grooves in these materials, their mechanical impedances are generally 
large enough compared with that of the recorders used to make the 
measurements of doubtful value as performance ratings. Curves 
Cl and C8 of this Fig. 4, taken with vertical reproducer C fitted with 
a diamond stylus, show the approximate range of frequency character- 
istic of materials of the cellulosic group (group 1} measured with 
vertical instruments. Curve Cl has been taken with material No. 1, 
and is the same material that was used for the lateral recording shown 



April, 1937] MATERIALS FOR DlSK RECORDING 421 

in curves Al and Bl of Fig. 1. The poorest material of group 1 for 
vertical recording, however, was not found to be the same as the 
poorest for lateral recording. 

Curve C5 shows the relative vertical cut frequency response char- 
acteristic of a material of group 3. The response is down about 
20 db. at 5000 cps. 

Curve C6 shows similar data for a material of group 4. The relative 
response is down about 10 db. at 5000 cps. 

Curve C9 shows similar data for a material of group 5. The 
relative response is down about 8 db. at 5000 cps. 

In taking these data on frequency characteristics, electrical equali- 
zation has not been used, in either recording or reproducing, in order 
to obtain the overall frequency response characteristic. It is possible 
to equalize within reasonable limits at either step and thereby to 
improve the frequency characteristic. When equalizing in recording, 
particularly at the higher frequencies, care must be taken to avoid 
excessive curvatures in the grooves for increasing recording levels, 
because these will give rise to severe distortion. When equalizing 
in reproducing, care must be taken to avoid excessive correction 
because this will increase surface noise. 

The value of frequency response measurements, although very 
great, does not give all the important facts that are necessary for 
quantitatively rating the performance of recording materials. Two 
other very important items are surface noise and distortion. 

Surface Noise. A tabulation of the data on surface noise for both 
lateral and vertical recording is shown in Table II. The noise 
measurements have been taken with a noise meter having a frequency 
weighting network and a damped meter of the volume indicator 
type. While this kind of noise measurement does not precisely 
check ear measurements, it is sufficiently accurate for the record 
measurements being taken here and has the advantage of being simple 
and quick, which is important because of the large number of measure- 
ments needed. The values are given in decibels, and are referred to 
a convenient arbitrary zero. The larger the db.- value given, the 
noisier the record. 

The measurements are again divided into five groups according 
to the material, and show a range of surface noise from the minimum 
value measured to the maximum measured, not, however, necessarily 
on materials of a single supplier. The range of values measured on 
the quietest record in each group is also given. 



422 



A. C. KELLER 



[J. S. M. P. E. 



The lateral measurements have been taken with reproducer A 
equipped with a steel stylus, except where noted, and with a record 
pressure of 50 grams. The vertical measurements have been taken 
with reproducer C equipped with a diamond stylus and at a record 
pressure of about 30 grams. The lateral and vertical noise levels are 
not exactly comparable because the vertical reproducer used had 
about twice the frequency range of the lateral reproducer. These 
noise measurements show that some of the records of group 1, as 
lateral records, are somewhat quieter than a commercial phonograph 

TABLE II 

Surface Noise Data in Decibels 



Material Group 


Lateral Recording 
Reproducer A, Steel Needle 
50-Gram Pressure 


Vertical Recording 
Reproducer C, Diamond Stylus 
30-Gram Pressure 


Range 


Best Record 


Range 


Best Record 


1 

2 

3 
4 
5 


10-45 
29-38* 
33-39 
25-32 
20-33 
27-38 


10-13 
29-32* 
33-36 
25-28 
20-27 
27-29 


11-52 
18-26 

25-31 
24-36 
28-37 


11-13 
18-26 

25-27 
24-30 
28-30 


Lateral cut phono- 
graph record 


23-40 








Vertical cut electrical 
transcription 






10-15 





* Measured with a fiber needle. 

(Note: The lateral and vertical record noise levels are not exactly comparable 
because of the wider frequency range of the vertical recording.) 

record. The surface noise level of vertical recording with materials 
of group 1 measured with a vertical reproducer are occasionally 
within a few db. of a quiet molded vertical cut transcription record. 
These measurements have been taken without equalization in the 
reproducer circuit to give a uniform frequency response characteristic. 
Such equalization for the characteristic measured on these materials 
will increase the measured noise levels. 

The measurements listed in Table II, excepting those of group 2, 
were all taken on records that were cut with the best steel recording 
styli tested, which closely approached the results obtained with 
sapphire styli. The measurements on records of group 2 were taken 



April, 1937] 



MATERIALS FOR DISK RECORDING 



423 



on records embossed with a diamond stylus. The noise levels on a 
given material with steel recording styli as received from several 
suppliers showed a variation of more than 10 db. 

In many cases, it has been difficult to obtain consistent data. 
Sometimes this has been due to variations in the material itself, and 
at other times to the effect of the recording materials upon the record- 
ing stylus. Some recording materials contained abrasives which 
caused rapid wear of the recording stylus. 

Wear Data. The record life has been measured in concentrically 
recorded unmodulated grooves with lateral and vertical instruments, 
and is shown in Table III. The record life has been taken as the 

TABLE m 
Record Life on Unmodulated Concentric Grooves 



Material Group 


Lateral Recording 
Reproducer A, 50-Gram 
Pressure, Steel Needle 


Vertical Recording 
Reproducer C, 30-Gram 
Pressure, Diamond Stylus 


1 


5->500t 


12- > 500 


2 


5-12* 


8-25 




2-4 




3 


22- > 600 


26- > 500 


4 


39-211 


32- > 500 


5 


4->500t 


235-305 


Lateral cut phono- 
graph record 


>500 




Vertical cut electrical 
transcription 




>500 



J High initial surface noise level. 
* Measured with a fiber needle. 

(Note: Vertical reproducer C has approximately twice the frequency range 
of lateral reproducer A .) 

number of times the groove is played for a 5-db. rise in noise level. 
Measurements were not carried beyond 500 playings. It has been 
assumed that the recorded levels on the records are such that a 5-db. 
rise in surface noise is reached before the record is regarded as worn 
out due to the wear on modulated grooves. This is not true for 
records made with excessive recording levels or with records repro- 
duced with instruments of high mechanical impedance. The repro- 
ducers and styli used for these tests were the same as those used in 
the surface noise tests. A number of measurements were obtained 



424 



A. C. KELLER 



[J. S. M. P. E. 



that indicated a record wi h a life greater than 500 playings. Some 
of these records, however, started with high initial noise level, and 
were so rough at the start that they showed little noise increment 
during test. These are specifically noted in the table. Omitting 
these records, many showed good wearing qualities adequate for 
many uses, even though not entirely comparable to commercial 
molded records. A life greater than 500 playings has been measured 
for commercial phonograph records. This is due to the low (50- 
gram) value of record pressure used. For a record pressure more 
commonly used, about 125 grams, this figure is reduced to 50 to 
100 playings. 

TABLE IV 

1000-Cycle Harmonic Distortion Data (Measurements Are Decibel Range Down 
from 1000-Cycle Component) 



Material Group 


Lateral 
Reproducer 


Recording 
A, Steel Needle 


Vertical 
Reproducer C, 


Recording 
Diamond Stylus 


2000 Cycles 


3000 Cycles 


2000 Cycles 


3000 Cycles 


1 


6-27 


14-36 


21-36 


28-39 


2 


12-25 


20-30 






3 


22-24 


27-32 


17-23 


Noisy 


4 


17-24 


27-35 


14-22 


27-34 


5 


17-26 


23-26 


15-21 


27-40 


Lateral cut phono- 










graph record 


20-30 


22-34 






Vertical cut elec- 










trical transcrip- 










tion 






28-34 


35-41 



Distortion Data. It is difficult to give a simple rating of distortion 
that will satisfy all conditions and will rate the various materials as 
a trained observer would for this factor if listening to speech or music. 
As a simple rough measurement data have been taken of the har- 
monics found in the reproduced output of a 1000-cycle recorded tone. 

The 1000-cycle data are tabulated for both lateral and vertical 
recording in Table IV. These data show that the materials are all 
somewhat inferior to molded records. The measurements have 
been corrected for frequency response characteristic. If equali- 
zation for frequency response had been used in recording, the dis- 
tortion would be greater than the tabulated measurements. The 
reproducers used for these measurements were reproducer B for the 
lateral and reproducer C for the vertical data. 



April, 1937] MATERIALS FOR DISK RECORDING 425 

Power Requirements. All materials tested, except aluminum and 
the pre-grooved material of group 4 shown in curve A 10 of Fig. 2, 
required power inputs to the recording instrument at 1000 cps. that 
were only a few decibels more than that required for wax recording. 
Aluminum, depending upon the source of supply, required about 
5 to 15 db. more recorder input power at 1000 cps. than did the con- 
trol wax recording. The group 4 material of Fig. 2 required about a 
20-db. increase in power input to the recorder. 

The increase of turntable power required has been found to be 
relatively moderate as compared with wax recording for most 
materials. Here again, aluminum required considerable increase in 
turntable power. No exact measurements of this have been made, 
but it is evident that the increase will be a function of the depth of 
the recorded groove. 

Processing. A few of the better records tested have been recorded 
and processed; that is, matrices have been electroplated from them 
and molded records made from these matrices. These few have 
shown an increase in surface noise in the processing of the order of 
10 db. Some of the other materials are not suitable for the present 
type of processing and would probably be re-recorded to wax if ma- 
trices were required. 

Fire Hazard. The materials of the cellulosic group (group 7) are 
most frequently supplied as coated metal disks, and are usually 
recorded by a cutting process. The shavings from all the materials 
tested of this group were found to be inflammable, due to the large 
amount of cellulose nitrate contained in them. The coated disks 
themselves are not readily inflammable because it is necessary to 
raise the metal core and the coating to kindling temperature in order 
for the record to burn, and this requires relatively long exposures at 
fairly high temperatures. The shavings from all other materials 
tested were found to be non-inflammable. 

Humidity Tests. All the materials tested were subjected to a 
humidity test at 95F. and 90 per cent relative humidity for 24 hours. 
The materials of group 1 showed a surface noise increase of from 
to 5 db. in this test. The coated metal disks showed inappreciable 
warping, but the thin disks of recording material alone showed con- 
siderable warping. The materials of groups 2, 3, and 4 showed little 
evidence of change in the humidity test except for warping of some 
materials of group 4, where thin disks of resin were used. The 
gelatin materials of group 5 were the most seriously affected in this 



426 A. C. KELLER 

test. The warping was generally very severe and noise measure- 
ments were impossible after the test, even for records where the 
material had been attached to metal disks. For the latter, severe 
wrinkling of the recording material occurred which caused regional 
separation from the metal base. 

Time did not permit an aging study of these materials which would 
correspond to storing them as recorded disks for long periods of time. 
For some uses very little change in the record is tolerable even over 
a period of years. 

Conclusion. The measurements show that considerable progress 
has been made in the field of direct recording materials in the past 
few years and that many of these disks are capable of reasonably 
satisfactory recording if used with care. They have been shown to 
have a life that should be adequate for many uses. The newer disks 
do not, in general, have the life of good molded records, so that their 
uses may often be regarded as complementary to those of molded 
records. The new records are thin and light in weight, and are 
therefore easily shipped and stored. Many of them are not easily 
damaged by ordinary handling, storing, or shipping. These newer 
direct recording materials will therefore probably greatly add to the 
rapidly growing use of recording. 



NEW MOTION PICTURE APPARATUS 

During the Conventions of the Society, symposiums on new motion picture appara- 
tus are held, in which various manufacturers of equipment describe and demonstrate 
their new products and developments. Some of this equipment is described in the 
following pages; the remainder will be published in subsequent issues of the Journal. 

THREE-WIRE D-C. SUPPLY FOR PROJECTION ARCS* 
C. C. DASH** 

Three-wire d-c. systems for projection room service provide a flexible and eco- 
nomic means for supplying current to the miscellaneous projection equipment 
usually found in the projection rooms of the larger theaters. Small theaters, 




IPKOJ 



FIG. 1. Connections for two type H. I. Transverters and panels. 



too, usually have spotlamps for special occasions, and the stereopticon also is a 
useful piece of projection equipment. 

The advent of the low-voltage d-c. projector lamps with their high efficiency 
when operated on d-c. sources particularly designed to supply current to such 
lamps has resulted in changes in generator equipment from high- to low-voltage 
generators. The auxiliary equipment could not be used on these low-voltage 

* Received September 21, 1936; presented at the Fall, 1936, Meeting at 
Rochester, N. Y. 
** Hertner Electric Co., Cleveland, Ohio. 

427 



428 



NEW MOTION PICTURE APPARATUS [J. S. M. p. E. 



sources, as it is outfitted with arcs that operate normally on 55 to 65 volts across 
the arc. 

The three-wire Edison system is admirably suited to meet these conditions in a 
projection room. The diagram in Fig. 1 shows how three-wire service has been 
provided by using two 42-volt generators. Each projector is operated on one 
of the motor-generator sets. The switching arrangement permits either pro- 
jector to be operated by either motor-generator set ; but when the stereopticon 
and spotlamps are to be used, both sets are operated and the d-c. outputs of the 
two are connected in series. It is obvious from the way in which the switching 
scheme is worked out that when only the projectors are being used, only one 
motor-generator set need be operated; and in the event of difficulty with either 



AlWU ARY 




FIG. 2. Generator connections for universal Trans- 
verter. 



of the motor-generators, current can still be supplied to the projectors from one 
motor-generator set. 

There are installations where it is not desirable to install duplicate motor- 
generator sets with two control panels and the necessary switching arrangement 
to take care of possible emergencies. A single-unit motor-generator set has been 
produced particularly for this service, in which two 42-volt generators are driven 
by one motor, the two generators being not only flat-compounded, but having 
auxiliary series fields so that the voltage of the generator is maintained constant 
regardless of changes of load and, consequently, of the speed of the driving motor 
when the load is placed upon the other generator. 

The motor of this particular type of motor-generator set is designed for high 
pull-out torque and high overall efficiency over a wide range of load. The genera- 
tors are identical. In order to maintain constant voltage over a wide range of 



April, 1937 



NEW MOTION PICTURE APPARATUS 



429 



load, it is desirable that the magnetic circuit of the generators be quite saturated. 
This provides stable operation with but little fluctuation with sudden changes of 
load, and with the large amount of space provided for shunt field windings the 
temperature rise in the shunt fields is extremely small. Current-densities in 
the armature and series field are extremely low so that the resistance drop is 
kept to an absolute minimum. A large number of commutator bars is used so as 
to improve the commutation and reduce the commutator ripple. By properly 




*t 

Eutreco T"-<E n-< HO.I 

FIG. 3. Performance curves of Universal Transverter. 



skewing the armature slots the commutator ripple is further reduced to a 
minimum. 

Fig. 2 shows the manner in which the series fields are connected into the circuit. 
Any load placed on No. 2 generator increases the excitation of generator No. 1 
sufficiently to compensate for the change of speed of the driving motor. Fig. 3 
shows the voltage regulation of this type of motor-generator set under various 
conditions ot load such as would be encountered in the projection room. 

When two motor-generator sets are used they are equipped with individual 
control panels, and the two machines are operated practically independently of 
each other. Any adjustment of the output voltage of one machine does not 
affect the output of the other. The panels may be equipped with ammeters 
or merely a voltmeter, depending upon whether or not the projectors are equipped 
with suitable meters. 

The panel equipment usually furnished with the two-generator motor-genera- 



430 



NEW MOTION PICTURE APPARATUS 



tor set has two independent field regulators, which are operated independently 
and which control the voltages of the generators independently so that the output 
voltage of either generator may be adjusted to the point at which the best opera- 
tion of the arc results. The voltmeter circuit is provided with three-way switches 
so connected into the circuit that the voltage across either generator or across 
the two in series may be read on the single voltmeter. 

The equipment was designed to meet the projection needs of some Midwest 









O 


b 








ZI^^-r- 

t . PL 


*{ j] SPOT 




"I 


SWITCH ^ 


O 


If 


* 




k 


^ i 

-4-2V 








*1 Sw 


TCM/ 


I 




















1 




























"V ^L 














- 






^ V H_ 


^n SPOT 


















,0-j 




i 


L, 

-FI 


1 \ ** 


feh 










SWITCH) 

N-l 


^7^ Voa*c>E REJOIN*; OF 


<5 




1 r 


s 










DOWN 


UP N" 1 &CNERPCTOR GNU/ 


ENtRM 


IB 1 


GEH 


rovroo 9 










DOWN 


DOWKI ho R.C*DtK^ 



FIG. 4. Connections for Universal Transverter and panel. 

theaters where 20- or 25-ampere generators were being used for reflector lamps 
with small carbons. To increase the current to 38 amperes for the S.R.A. carbons 
it was necessary to buy new generators, and in order to take care of the possible 
introduction of the Suprex lamp at a later date this particular type of generator 
was installed. 

The generators are built in various capacities as high as 200 amperes on 
continuous duty. The momentary service, or momentary overload duty, dur- 
ing a change-over period would be 300 amperes per generator at 84 volts. 



COMMITTEES 

of the 
SOCIETY OF MOTION PICTURE ENGINEERS 

(Correct to March 20th; additional appointments or changes may be made at 
any time during the year as necessity or expediency may require.} 



L. W. DAVEE 
H. GRIFFIN 



ADMISSIONS AND TRANSFERS 

G. FRIEDL, JR., Chairman 
S. HARRIS 
D. E. HYNDMAN 



P. J. LARSEN 
M. W. PALMER 



A. N. GOLDSMITH 
A. C. HARDY 



BOARD OF EDITORS 

J. I. CRABTREE, Chairman 

L. A. JONES H. G. KNOX 

E. W. KELLOGG T. E. SHEA 



COLLEGE COURSE IN TECHNICAL MOTION PICTURE EDUCATION 
T. E. SHEA, Chairman 



A. N. GOLDSMITH 



L. A. JONES 



W. H. CARSON 
O. O. CECCARTNI 



COLOR 

J. A. BALL, Chairman 
C. H. DUNNING 
R. M. EVANS 



A. M. GUNDELFINGER 

H. W. MOYSE 



H. GRIFFIN 

J. H. KURLANDER 



CONVENTION 

W. C. KUNZMANN, Chairman 

P. MOLE K. F. MORGAN 

M. W. PALMER 



H. WARNCKE 



EUROPEAN ADVISORY COMMITTEE 

J. VAN BREUKELEN, Chairman 

F. H. HOTCHKISS I. D. WRATTEN 



O. C. BINDER 
A. S. DICKINSON 
G. K. HADDOW 



EXCHANGE PRACTICE 

A. W. SCHWALBERG, Chairman 



K. C. KAUFMAN 
J. S. MACLEOD 
H. A. MBRSAY 



N. F. OAKLEY 
H. RUBIN 
J. H. SPRAY 



431 



432 



COMMITTEES OF THE SOCIETY 
HISTORICAL 



[J. S. M. P. E. 



J. E. ABBOTT 
T. ARM AT 



O. B. DEPUE 



A. N. GOLDSMITH 



A. C. HARDY 



E. THEISEN, Chairman 
G. A. CHAMBERS 
W. CLARK 

MUSEUM 

(Western) 
E. THEISEN, Chairman 

J. A. DUBRAY 

HONORARY MEMBERSHIP 
J. I. CRABTREE, Chairman 

JOURNAL AWARD 

E. A. WILLIFORD, Chairman 
E. HUSE 



G. E. MATTHEWS 
T. RAMSAYE 



A. REEVES 



E. A. WILLIFORD 



G. F. RACKETT 



L. A. BONN 
R. M. EVANS 
G. GIBSON 
E. HUSE 



D. P. BEAN 
F. E. CARLSON 
W. B. COOK 
H. A. DEVRY 



C. N. BATSEL 
L. N. BUSCH 
O. O. CECCARINI 
A. A. COOK 
L. J. J. DIDIEE 



J. E. ABBOTT 
J. I. CRABTREE 
A. S. DICKINSON 



LABORATORY PRACTICE 

D. E. HYNDMAN, Chairman 

T. M. INGMAN H. W. MOYSE 

C. L. LOOTENS J. M. NlCKOLAUS 

R. F. MITCHELL W. A. SCHMIDT 

J. H. SPRAY 

NON-THEATRICAL EQUIPMENT 

R. F. MITCHELL, Chairman 

E. C. FRITTS R. C. HOLSLAG 

H. GRIFFIN J. H. KURLANDER 

J. A. HAMMOND 



A. SHAPIRO 
A. F. VICTOR 



PAPERS 



G. E. MATTHEWS, Chairman 

C. FLANNAGAN T. E. SHEA 

M. E. GILLETTE 

E. W. KELLOGG 

R. F. MITCHELL 

W. A. MUELLER 

E. C. RICHARDSON 



P. R. VON SCHROTT 

H. C. SILENT 
H. G. TASKER 
I. D. WRATTEN 



PRESERVATION OF FILM 

J. G. BRADLEY, Chairman 
R. EVANS 
M. E. GILLETTE 
C. L. GREGORY 



T. RAMSAYE 
V. B. SEASB 
W. A. SCHMIDT . 



April, 1937] 



COMMITTEES OF THE SOCIETY 



433 



MEMBERSHIP AND SUBSCRIPTION 



Alabama 


Michigan 


Ohio 


P. A. KING 


J. F. STRICKLER 


C. C. DASH 






R. H. GILES 


California 


Minnesota 


V. C. WELMAN 


J. O. AALBERG 


C. L. GREENE 




C. W. HANDLEY 


R. H. RAY 


Pennsylvania 


E. HUSE 




H. BLOOMBERG 


R. H. McCULLOUGH 


Missouri 


A. GOODMAN 


G. A. MITCHELL 


J. S. COPLEY 


I. SAMUELS 


P. MOLE 






K. F. MORGAN 


New York 


Texas 


W. A. MUELLER 


A. BECKER 


H. H. FRASCH 


H. G. TASKER 


F. E. CAHILL 






A. A. COOK 


District of Columbia 


Georgia 


A. S. DICKINSON 


H. T. COWLING 


N. WEIL 


J. J. FINN 


R. EVANS 




J. FRANK, JR. 


N. D. GOLDEN 


Illinois 


S. HARRIS 


F. J. STORTY 


H. A. DEVRY 


D. E. HYNDMAN 




B. J. KLEERUP 


W. H. INGRAM 


Travelling 


S. A. LUKES 


O. E. MILLER 


E. AUGER 


C. G. OLLINGER 


F. H. RICHARDSON 


C. BRENKERT 


J. M. SCHAEFER 


P. D. RIES 


F. HOHMEISTER 


J. H. TOLER 


C. J. STAUD 


W. C. KUNZMANN 




L. M. TOWNSEND 


D. McRAE 


Massachusetts 


J. S. WARD 


O. F. NEU 


J. S. ClFRE 




H. H. STRONG 


S. SUMNER 


Japan 




A. B. WEST 


T. NAGASE 


Germany 




Y. OWAWA 


W. F. BIELICKB 


Australia 






H. C. PARRISH 


New Zealand 


Hawaii 




C. BANKS 


L. LA CHAPELLB 


Austria 






P. R. VON SCHROTT 


England 


Holland 




W. F. GARLING 


J. VAN BREUKELEN 


Canada 


R. G. LlNDERMAN 




F. C. BADGLEY 


E. McMASTER 


India 


G. H. BATTLE 


R. TERRANEAU 


G. D. LAL 


C. A. DENTELBECK 




H. S. MEHTA 


B. E. NORRISH 


France 


M. L. MISTRY 




L. J. DIDIEE 




China 


L. G. EGROT 


Russia 


R. E. O'BOLGER 


F. H. HOTCHKISS 


E. G. JACHONTOW 



434 



COMMITTEES OF THE SOCIETY 
PROGRESS 



[J. S. M. P. E. 



L. N. BUSCH 
G. A. CHAMBERS 
A. A. COOK 



J. G. FRAYNE, Chairman 
R. M. CORBIN 
R. E. FARNHAM 
W. LEAHY 



G. E. MATTHEWS 
V. E. MILLER 
G. WORRALL 



J. I. CRABTREE 



J. O. BAKER 
T. C. BARROWS 
F. E. CAHILL 
J. R. CAMERON 
A. A. COOK 
J. K. ELDERKIN 
J. J. FINN 
R. R. FRENCH 
E. R. GEIB 



PROGRESS AWARD 

A. N. GOLDSMITH, Chairman 

M. C. BATSEL R. M. EVANS 



PROJECTION PRACTICE 

H. RUBIN, Chairman 
A. N. GOLDSMITH 
A. GOODMAN 
H. GRIFFIN 
S. HARRIS 
J. J. HOPKINS 

C. F. HORSTMAN 

D. E. HYNDMAN 
P. A. McGuiRB 
R. MIEHLING 



E. R. MORIN 

M. D. O'BRIEN 

G. F. RACKETT 

F. H. RICHARDSON 

B. SCHLANGER 

C. TUTTLE 

J. S. WARD 
V. A. WELMAN 
A. T. WILLIAMS 



J. R. CAMERON 
J. J. FINN 



PUBLICITY 

W. WHITMORE, Chairman 
S. HARRIS 
G. E. MATTHEWS 
W. A. MUELLER 



P. A. McGuiRE 
F. H. RICHARDSON 



P. ARNOLD 
M. C. BATSEL 
F. C. BADGLEY 
L. N. BUSCH 
A. CHORINE 
A. COTTET 
L. DE FEO 
A. C. DOWNES 

J. A. DUBRAY 

P. H. EVANS 



STANDARDS 

E. K. CARVER, Chairman 
R. E. FARNHAM 
C. L. FARRAND 
G. FRIEDL, JR. 
H. GRIFFIN 
A. C. HARDY 
R. C. HUBBARD 
E. HUSE 
C. L. LOOTENS 
K. F. MORGAN 
T. NAGASE 



N. F. OAKLEY 
G. F. RACKETT 
W. B. RAYTON 
C. N. REIFSTECK 
H. RUBIN 

0. SANDVIK 
H. B. SANTEE 
J. L. SPENCE 

J. VAN BREUKELEN 

1. D. WRATTEN 



STUDIO LIGHTING 

R. E. FARNHAM, Chairman 
W. C. KUNZMANN V. E. MILLER E. C. RICHARDSON 

J. H. KURLANDER M. W. PALMER F. WALLER 

G. F. RACKETT 



April, 1937] COMMITTEES OF THE SOCIETY 435 

SECTIONS OF THE SOCIETY 

(Atlantic Coast) 

G. FRIEDL, JR., Chairman 

L. W. DAVEE, Past-Chairman M. C. BATSBL, Manager 

D. E. HYNDMAN, Sec.-Trcas. H. GRIFFIN, Manager 

(Mid-West) 

C. H. STONE, Chairman 

R. F. MITCHELL, Past-Chairman O. B. DEPUE, Manager 

S. A. LUKES, Sec.-Treas. B. E. STECHBART, Manager 

(Pacific Coast) 

K. F. MORGAN, Chairman 

H. G. TASKER, Past-Chairman J. O. AALBERG, Manager 

G. A. CHAMBERS, Sec.-Treas. H. W. MOYSE, Manager 



SPRING, 1937, CONVENTION 
SOCIETY OF MOTION PICTURE ENGINEERS 

HOLLYWOOD-ROOSEVELT HOTEL 
HOLLYWOOD, CALIF. 
MAY 24th-28th, INCL. 

Officers and Committees in Charge 

W. C. KUNZMANN, Convention Vice-President 

J. I. CRABTREE, Editorial Vice-President 

H. G. TASKER, Past-President 

G. F. RACKETT, Executive Vice-President 

K. F. MORGAN, Chairman, Pacific Coast Section 

G. E. MATTHEWS, Chairman, Papers Committee 

Local Papers Committee 

W. A. MUELLER, Chairman 
L. A. AICHOLTZ, Secretary 

C. N. BATSEL E. C. RICHARDSON H. C. SILENT 

O. O. CECCARINI H. G. TASKER 

Local Arrangements and Reception Committee 

P. MOLE, Chairman 

E. HUSE K. F. MORGAN H. G. TASKER 

J. O. AALBERG C. W. HANDLEY G. F. RACKETT 

C. N. BATSEL R. H. McCuLLOUGH H. W. MOYSE 

G. A. CHAMBERS G. S. MITCHELL W. J. QUINLAN 

Information and Registration 

W. C. KUNZMANN, Chairman 
E. R. GEIB S. HARRIS C. W. HANDLEY 

Ladies' Reception Committee 

MRS. K. F. MORGAN and MRS. P. MOLE, Hostesses 

assisted by 

MRS. F. C. COATES MRS. E. HUSE MRS. E. C. RICHARDSON 

MRS. C. W. HANDLEY MRS. W. J. QUINLAN MRS. H. G. TASKER 

MRS. G. F. RACKETT 
436 



SPRING CONVENTION 437 

Banquet 

E. HUSE, Chairman 

W. C. KUNZMANN K. F. MORGAN G. F. RACKETT 

P. MOLE W. J. QUINLAN H. G. TASKER 

Projection Committee 

H. GRIFFIN, Chairman 

J. O. AALBERG J. FRANK, JR. C. W. HANDLEY 

L. E. CLARK J. G. FRAYNE R. H. McCuLLOUGH 

G. M. GROSJEAN 
Officers and Members of Los Angeles Local No. 150, I.A.T.S.E. 

Hotel Accommodations Committee 

G. F. RACKETT, Chairman 
E. HUSE K. F. MORGAN 

W. C. KUNZMANN H. G. TASKER 

H. C. SILENT 

Transportation Committee 

C. W. HANDLEY, Chairman 
G. A. CHAMBERS S. HARRIS 

H. GRIFFIN F. E. JAMES 

Publicity 

W. WHITMORE, Chairman 
J. J. FINN S. HARRIS 

W. GREENE G. E. MATTHEWS 

W. A. MUELLER 

Membership 

E. R. GEIB, Chairman 
G. A. CHAMBERS W. GREENE S. HARRIS 

Headquarters 

Headquarters of the Convention will be the Hollywood-Roosevelt Hotel, where 
excellent accommodations are assured. A reception suite will be provided for the 
Ladies' Committee, and an excellent program of entertainment will be arranged 
for the ladies who attend the Convention. 

Special hotel rates, guaranteed to SMPE delegates, European plan, will be as 
follows: 

One person, room and bath $ 3 . 50 

Two persons, double bed and bath 5.00 
Two persons, twin beds and bath 6 . 00 
Parlor suite and bath, 1 person 9 . 00 

Parlor suite and bath, 2 persons 12 . 00 



438 SPRING CONVENTION [j. S. M. p. E. 

Room reservation cards will be mailed to the membership of the Society in the 
near future, and everyone who plans to attend the Convention should return his 
card to the Hotel promptly in order to be assured of satisfactory accommodations. 
Special garage rates will be provided for SMPE delegates who motor to the 
Convention. 

Railroad Fares 

The dates of the Convention have been chosen in order that delegates may avail 
themselves of the summer tourists' rates, which go into effect May 15th. The 
following table lists the railroad fares and Pullman charges: 

Railroad 

Fare Pullman 

City (round trip) (one way) 

Washington $120.75 $20.50 

Chicago 86.00 15.75 

Boston 132.80 22.25 

Detroit 98.30 18.00 

New York 126.90 21.75 

Rochester 112.50 19.25 

Cleveland 101.35 18.00 

Philadelphia 122.85 21.00 

Pittsburgh 107.10 18.75 

The railroad fares given above are for round trips, forty-five day limits. Ar- 
rangements may be made with the railroads to take different routes going and 
coming, if so desired, but once the choice is made it must be adhered to, as changes 
in the itinerary may be effected only with considerable difficulty and formality. 
Delegates should consult their local passenger agents as to schedules, rates, and 
stop-over privileges. 

New streamlined trains will be operating from Chicago to Los Angeles and San 
Francisco, making the trip to Los Angeles in 39 hours. Special fares are levied on 
these trains. 

Technical Sessions 

The Hollywood meeting always offers our membership a rare opportunity to 
become better acquainted with the studio technicians and production problems. 
Accordingly, arrangements are being made to hold two evening sessions at two 
of the studios. The Monday evening session will be devoted to a practical 
demonstration on a studio set of the function of the various personnel units who 
contribute to making a picture. On Tuesday evening arrangements are being 
made to demonstrate outstanding examples of sound recording and color photog- 
raphy, special effects, and picture quality. Also tentatively scheduled for this 
evening is a demonstration of stereophonic sound reproduction by Mr. Douglas 
Shearer. 

The Academy of Motion Picture Arts & Sciences is arranging a session by 
leading Academy members, and reports will also be made of the work of the 
various Academy committees. 

The general technical sessions will include papers on production problems, 
studio design and organization, push-pull recording and reproduction, critically 



April, 1937] SPRING CONVENTION 439 

damped filters, electrical engineering problems and equipment for studios, film 
storage, density measurements, and other pertinent subjects. 

An endeavor is being made to schedule a symposium on the production of 
color stills. A new color process will be described and demonstrated, and papers 
on lighting for color pictures have been promised. 

A large number of interesting papers are promised for the Apparatus Sym- 
posium. 

The Local Papers Committee under the chairmanship of Mr. William A. 
Mueller and with Lawrence Aicholtz as secretary is collaborating closely with 
the General Papers Committee in arranging the details of the program. Other 
members of this committee are : C. N. Batsel, O. O. Ceccarini, E. C. Richardson, 
H. C. Silent, and H. G. Tasker. 

Complete details of the program will be published in the May issue of the 
JOURNAL. 

Semi- Annual Banquet 

The Semi- Annual Banquet of the Society will be held at the Hotel on Wednes- 
day, May 26th. Addresses will be delivered by prominent members of the indus- 
try, followed by dancing and entertainment. Tables reserved for 8, 10, or 12 per- 
sons; tickets obtainable at the registration desk. 

Inspection Tours and Diversions 

Arrangements are under way to visit one or more of the prominent Hollywood 
studios, and passes will be available to registered members to several Hollywood 
motion picture theaters. Arrangements may be made for golfing and for special 
trips to points of interest in and about Hollywood. 

Ladies' Program 

An especially attractive program for the ladies attending the Convention is be- 
ing arranged by Mrs. K. F. Morgan and Mrs. P. Mole, hostesses, and their Ladies' 
Committee. A suite will be provided in the Hotel, where the ladies will register 
and meet for the various events upon their program. Further details will be pub- 
lished in a succeeding issue of the JOURNAL. 

Points of Interest 

En route: Boulder Dam, Las Vegas, Nevada; and the various National Parks. 
Hollywood and vicinity: Beautiful Catalina Island; Zeiss Planetarium; Mt. 
Wilson Observatory; Lookout Point, on Lookout Mountain; Huntington Li- 
brary and Art Gallery (by appointment only); Palm Springs, Calif.; Beaches at 
Ocean Park and Venice, Calif.; famous old Spanish missions; Los Angeles Mu- 
seum (housing the SMPE motion picture exhibit); Mexican village and street. 
Los Angeles. 

TENTATIVE PROGRAM 
Monday, May 24th 
10:00 a.m. Blossom Room 
Registration 
Society Business 
Committee Reports 
Technical Papers Program 



440 SPRING CONVENTION [J. S. M. p. E. 

12:30 p.m. Florentine Room 

Informal Get-Together Luncheon for members, their families, and 
guests. Brief addresses by several prominent members of the 
industry. 
2 :00 p.m. Blossom Room 

Technical Papers Program. 
8:00 p.m. (To be announced later.) 

Tuesday, May 25th 

10:00 a.m. Blossom Room 

Technical Papers Program 
2:00 p.m. (To be announced later.) 
8:00 p.m. Blossom Room 

Technical Papers Program 

Wednesday, May 26th 

10:00 a.m. Blossom Room 

Technical Papers Program 
2:00 p.m. (To be announced later.) 
8:00 p.m. Blossom Room 

Semi- Annual Banquet and Dance of the SMPE; Addresses by 
eminent members of the industry; dancing and entertainment. 

Thursday, May 27th 

10:00 a.m. Open morning 
2:00 p.m. Blossom Room 

Technical Papers Program 
8:00 p.m. Blossom Room 

Technical Papers Program 

Friday, May 28th 

10:00 a.m. Blossom Room 

Technical Papers Program 
2:00 p.m. Blossom Room 

Technical Papers Program 

Open Forum 

Adjournment of the Convention 

NOTE: All technical sessions will be held in the Blossom 
Room of the Hollywood-Roosevelt Hotel. There will be no 
public exhibit of apparatus in the Hotel; although members 
registered in the Hotel will, of course, be privileged to display 
any equipment they wish in their own rooms. 



SOCIETY ANNOUNCEMENTS 

ATLANTIC COAST SECTION 

At a meeting held on January 20th at the Hotel Pennsylvania, New York, 
N. Y., a symposium of three presentations was held on the subject of lighting. 
Mr. G. Mili, of the Westinghouse Electric & Manufacturing Company, Bloom- 
field, N. J., presented a paper on the subject of "Light and Light Control in 
Photography," followed by a discussion of "New Developments in Lamps for 
Photography," by J. H. Kurlander of the same company. The papers were 
accompanied by a demonstration of lighting equipment employing incandescent 
lamps used in studio practice. In addition, Mr. F. E. Eldredge, also of the 
Westinghouse Electric & Manufacturing Company, re-presented the paper, "A 
Demonstration Triode Tube," written jointly with H. F. Dart, previously pre- 
sented at the Rochester Convention last October. 

At the meeting held on February 25th at the studio of Movietonews, Inc., Mr. 
H.I. Reiskind presented a paper describing the new RCA single-channel recording 
and re-recording installation recently completed at the Movietonews studio. 
Following the presentation, members attending the meeting were privileged to 
inspect the equipment and listen to several samples of recent recordings made 
on the system. 

Both meetings were very well attended, in the latter case to the extent of more 
than two hundred members and guests, and considerable discussion followed 
the presentations. 

MID-WEST SECTION 

On March llth, at the auditorium of Bell and Howell Co., Chicago, a paper 
entitled "Some Aspects of Motion Picture Photography" was presented by 
D. E. Hyndman, of the Eastern Division Motion Picture Film Department of 
the Eastman Kodak Company. 

The next meeting of the Section is scheduled for April 8th. 

STANDARDS COMMITTEE 

At a meeting held on February 18th at the General Office of the Society, 
further consideration was given to the revision of the standards drawings as a 
consequence of the actions taken some months ago at the meeting of the Inter- 
national Standards Association at Budapest, and of various proposals made 
in the interim for the purpose of achieving greater unanimity between the stand- 
ards of the S.M.P.E. and those of various European standardizing organizations. 

The revision has been practically completed, and the Committee will probably 
be voting upon adopting them about the time this issue goes to press, after which 
the Board of Governors will take action upon submitting the revisions to the 
American Standards Association for formal adoption as American national 
standards. 

441 



442 



SOCIETY ANNOUNCEMENTS 



[J. S. M. P. E. 



SECTIONAL COMMITTEE ON MOTION PICTURES, ASA 

On March 10th, a meeting of Technical Sub-Committee No. 2 Optical, 
under the chairmanship of J. O. Baker, was held at the General Office of the 
Society, the purpose of this meeting being primarily to establish the scope of 
activity of the Sub-Committee and to formulate an agenda for the coming 
season. 

The Sub-Committee plans to investigate the standards of the various scientific 
and engineering societies for the purpose of determining whether any such stand- 
ards may be suitable for adoption as American motion picture standards. 



ADMISSIONS COMMITTEE 

At a recent meeting of the Admissions Committee, at the General Office of 
the Society, the following applicants for membership were admitted to the 
Associate grade : 



BECKWITH, R. F. 
Recordak Corp. 
235 W. 23d St. 
New York, N. Y. 
BINDER, M. 
42 E. 28th St. 

New York, N. Y. 
BISHOP, M. 
40 Grey St. 
Gisborne, N. Z. 

BORRAS, A. 

Cinefilm Laboratories 

Large St. & Solly Ave. 

Philadelphia, Pa. 
DARLING, C. W. 
3412 Addington Ave. 

Montreal, P. Q. 

Canada. 
DENNISON, L. I. 
Patterson & Dennison, Inc. 

6210 Greenfield Ave. 

West Allis, Wis. 
GOULD, M. E. 

Photoelectric Business Machines, Inc. 

74 Trinity Place 

New York, N. Y. 

HUMBERSTONE, F. G. 

39, Highdown 
Worcester Park 
Surrey, England. 



KERDEL, O. 

Llaguno a Bolero 44 
Caracas, Venezuela. 

KLINEDINST, M. S. 
395 Warburton Ave. 
Yonkers, N. Y. 

KOJIMA, E. 

The K. S. Talkie Seisakusho 
Kyobunkan Bldg., 3 chome 
Ginza, Tokyo, Japan. 

MARASCO, B. 
Via Melchiorra Gioia 121 
Milan, Italy. 

MEEK, C. P. 
4165 Ellis Ave. 
Chicago, 111. 

MERSAY, H. A. 

20th Century-Fox Film Corp. 
444 W. 56th St. 
New York, N. Y. 

NAKAJIMA, Z. 

D. Nagase & Co., Ltd. 
7, Itachibori Minamidori 
1 chome, Osaka, Japan. 

OGURA, J. 

D. Nagase & Co., Ltd. 
7, Itachibori Minamidori 
1 chome, Osaka, Japan. 



April, 1937] SOCIETY ANNOUNCEMENTS 443 

RAMSEY, A. B. ULMER, A. R. 

Ramsey Tower Bldg. 64 Shelby St. 

Oklahoma City, Okla. Dumont, N. J. 

ROBACH, M. WALTERS, W. H. 

c/o Film Daily Walters Electric Co. 

1501 Broadway 739 Third Ave. 

New York, N. Y. New York, N. Y. 

WHEATLEY, R. T. 

Tofo t' < A T> RCA Photophone, Ltd. 

421 S. Roxford Road Blythswood Sq. 

Syracuse, N. Y. Glasgow ^^ 

STOCKER, A. J. WOLAK, A. J. 

551 Fifth Ave. 114 Belvidere St. 

New York, N. Y. Waukegan, 111. 

In addition, the following applicants have been admitted by vote of the Board 
of Governors to the Active grade : 

HOPPER, F. L., PEMBBRTON, H. L., 

599 S. Hudson Ave., 9 Place des Martyrs, 

Pasadena, Calif. Brussels, Belgium. 

KOENIG, W., REISKIND, H. I., 

Universal Pictures, RCA Manufacturing Co., 

Universal City, Calif. Camden, N. J. 



The Society regrets to announce the death of 

MAX MAYER 
September 15, 1936 

Mr. Mayer was one of the founder members of the Society, and held office in 
the Society as Vice-President from 1918 to 1921, inclusive, and was elected member 
of the Board during 1928 and 1929. He continued as a member of the Society 
up to 1931, at about which time his health began to fail. Mr. Mayer's activities 
were concerned primarily with arc lamp lighting, although prior to 1912 he had 
been engaged in much experimental work with x-rays and the fluoroscope. In 
1915 he became interested in the Prizma process of color cinematography, and 
for some years was associated with the M. J. Wohl Company of New York, 
manufacturers of arc lamps. 



The Society regrets to announce the death of 

J. P. SKELLY 
March 8, 1937 



STANDARD S. M. P. E. 

VISUAL AND SOUND TEST REELS 

Prepared under the Supervision 

OF THE 
PROJECTION PRACTICE COMMITTEE 

OF THE 
SOCIETY OF MOTION PICTURE ENGINEERS 

<! & ;> 

Two reels, each approximately 500 feet long, of specially pre- 
pared film, designed to be used as a precision instrument in 
theaters, review rooms, exchanges, laboratories, and the like 
for testing the performance of projectors. The visual section 
includes special targets with the aid of which travel-ghost, 
lens aberration, definition, and film weave may be detected 
and corrected. The sound section includes recordings of 
various kinds of music and voice, in addition to constant- 
frequency, constant-amplitude recordings which may be used 
for testing the quality of reproduction, the frequency range 
of the reproducer, the presence of flutter and 60-cycle or 96- 
cycle modulation, and the adjustment of the sound -track. 
Reels sold complete only (no short sections). 

PRICE $37.50 FOR EACH SECTION, 
INCLUDING INSTRUCTIONS 

(Shipped to any point in the United States) 

Address the 

SOCIETY OF MOTION PICTURE ENGINEERS 

HOTEL PENNSYLVANIA 
NEW YORK, N. Y. 



JOURNAL 

OF THE SOCIETY OF 

MOTION PICTURE ENGINEERS 

Volume XXVIII MAY, 1937 Numbers 



CONTENTS 

Page 

A Brief Survey of the Physics and Technology of the Berthon- 
Siemens Color Process E. GRETENER 447 

Recent Developments in Magnetic Sound Recording 

S. J. BEGUN 464 

Iconoscopes and Kinescopes in Television . .-. . .V. K. ZWORYKIN 473 

New Motion Picture Apparatus 

A Single-Channel Recording and Re- Recording System 

H.I. REISKIND 498 

The RCA Sound Recording System 

M. C. BATSEL AND E. W. KELLOGG 507 

Book Reviews 534 

Open Forum 536 

Officers and Governors of the Society 538 

Committees of the Society 541 

Spring, 1937, Convention; Hollywood, Calif., May 24th-28th, 
Inclusive : 

General 546 

Tentative Program 550 

Abstracts of Papers and Presentations 557 

Society Announcements 580 



JOURNAL 

OF THE SOCIETY OF 

MOTION PICTURE ENGINEERS 



SYLVAN HARRIS, EDITOR 

Board of Editors 
J. I. CRABTREE, Chairman 

A. N. GOLDSMITH L. A. JONES H. G. KNOX 

A. C. HARDY E. W. KELLOGG T. E. SHEA 



Subscription to non-members, $8.00 per annum ; to members, $5.00 per annum, 
included in their annual membership dues; single copies, $1.00. A discount 
on subscriptions or single copies of 15 per cent is allowed to accredited agencies. 
Order from the Society of Motion Picture Engineers, Inc., 20th and Northampton 
Sts., Easton, Pa., or Hotel Pennsylvania, New York, N. Y. 
Published monthly at Easton, Pa., by the Society of Motion Picture Engineers. 

Publication Ofl\ce, 20th & Northampton Sts., Easton, Pa. 
General and Editorial Office, Hotel Pennsylvania, New York, N. Y. 

West-Coast Office, Suite 226, Equitable Bldg., Hollywood, Calif. 
Entered as second class matter January 15, 1930, at the Post Office at Easton, 
Pa., under the Act of March 3, 1879. Copyrighted, 1937, by the Society of 
Motion Picture Engineers, Inc. 

Papers appearing in this Journal may be reprinted, abstracted, or abridged 
provided credit is given to the Journal of the Society of Motion Picture Engineers 
and to the author, or authors, of the papers hi question. Exact reference as to 
the volume, number, and page of the Journal must be given. The Society is 
not responsible for statements made by authors. 



OFFICERS OF THE SOCIETY 

President: S. K. WOLF, 250 W. 57th St., New York, N. Y. 
Past-President: H. G. TASKER, Universal City, Calif. 

Executive Vice-President, G. F. RACKETT, 823 N. Seward St., Hollywood, Calif. 
Engineering Vice-President: L. A. JONES, Kodak Park, Rochester, N. Y. 
Editorial Vice-President: J. I. CRABTREE, Kodak Park, Rochester, N. Y. 
Financial Vice-President: O. M. GLUNT, 463 West St., New York, N. Y. 
Convention Vice-President: W. C. KUNZMANN, Box 6087, Cleveland, Ohio. 
Secretary: J. FRANK, JR., 90 Gold St., New York, N. Y. 
Treasurer: L. W. DAVEE, 250 W. 57th St., New York, N. Y. 

GOVERNORS 

M. C. BATSEL, Front and Market Sts., Camden, N. J. 

A. S. DICKINSON, 28 W. 44th St., New York, N. Y. 

G. FRIEDL, JR., 250 W. 57th St., New York, N. Y. 

A. N. GOLDSMITH, 444 Madison Ave., New York, N. Y. 

H. GRIFFIN, 90 Gold St., New York, N. Y. 

A. C. HARDY, Massachusetts Institute of Technology, Cambridge, Mass. 

K. F. MORGAN, 7046 Hollywood Blvd., Los Angeles, Calif. 

C. H. STONE, 205 W. Wacker Drive, Chicago, 111. 



See p. 541 for Technical Committees 



A BRIEF SURVEY OF THE PHYSICS AND TECHNOLOGY 
OF THE BERTHON-SIEMENS COLOR PROCESS* 



E. GRETENER** 

Summary. An analysis is made of the advantages and limitations of the lenticu- 
lar color-film system over that of the screen-film method as well as processes using three 
separate images. Exposure of lenticular films requires a perfect system of objectives 
with uniform optical filter position. Entirely new types of negative emulsions of 
high resolving power were devised to overcome irradiation effects. 

Methods of printing lenticular films are described. Objectives of small aperture 
were developed to offset color fringing resulting when large-aperture objectives were 
used. The three separation images are registered by a mirror system. 

New arc lamps of very high efficiency are described for use in projection. They 
operate with horizontally arranged square carbons and magnetic stabilization. The 
luminous gas portion extends in front of the positive carbon and can radiate laterally 
without obstruction. With 60 amperes the intrinsic brilliance in the crater was found 
to be SOO candles per sq. mm. 

A further increase of screen brightness was realized by the use of a special projec- 
tion screen which confined the reflected light within the needed angle. Metal sheets 
are used into which one million small concave mirrors are rolled per square meter. 
On the basis of conditions prevailing in the German theaters, an increase of screen 
brightness by a factor of 3 can be attained with the use of the new screen. 

In the past few years in Europe the lens-screen method for color 
films has been further developed. The most active group worked 
with the inventor, Rodolphe Berthon. Their interests are embodied 
in Opticolor A.G., Glarus, Switzerland. At the request of this com- 
pany, Siemens & Halske A. G. (Berlin), in connection with other firms, 
developed the method into an industrial process. It has received 
the name Berthon-Siemens Color-Film Process. During the 1936 
Olympic games a short film, made by this process, was shown pub- 
licly in Berlin. It is believed that this represents the first time in 
the world that a lens-screen color-film (35-mm.) has been shown suc- 
cessfully on a motion picture theater program. Additional films are 
being made. 

* Received September 29, 1936; presented by title at the Fall, 1936, 
Meeting at Rochester, N. Y. 

** Siemens & Halske A. G., Berlin-Siemensstadt, Germany. 

447 



448 E. GRETENER [J. S. M. P. E. 

INTRODUCTION 

(1) Color Distortion 

Making satisfactory color motion pictures is largely a problem in 
reproducing the impressions produced upon the senses when an ob- 
ject is viewed. Equivalence between the original colored object and 
its reproduction results when definite properties of a mixture of wave- 
lengths for a color print are carried through the entire process. It 
is well known that a color can be characterized clearly by a statement 
of three numbers, designated as tri-stimulus values. Any color can 
be matched by an additive mixture of suitable amounts of three 
fundamental colors, red, green, and blue. The amounts of the funda- 
mental colors necessary to match a sample color are called the tri- 
stimulus values of the color. A primary color coordinate system is 
obtained in terms of these tri-stimulus values. Specification of domi- 
nant wavelength, purity, and brightness is thereby much simplified. 
These quantities may be derived in a simple manner for any color 
from the coordinates of the color in the tri-stimulus coordinate system. 

The performance of color-film processes can be judged at present 
by the color distortion. The way in which the color coordinates of 
the reproduction are related to those of the original object is deter- 
mined. Indistinguishability between the reproduction and the original 
is possible only with identity of the tri-stimulus values. In general, 
there are certain distortions in dominant wavelength, purity, and 
brightness, which may be considered analogous to the sound distor- 
tions occurring in sound-films. However, alteration of the color 
coordinates by reproduction does not provide a measure of the 
physiological effect of the print. The relations are the same here as 
in sound-film in respect to which, likewise, no direct conclusion as to 
quality of tone reproduction can be drawn from a statement of the 
frequency characteristic or the noise factor. It is a case of learning 
to determine which physical color distortion can be permitted with- 
out giving rise to an unnatural appearance of the color print. 

A physiological measurement of the difference between colors has 
been attempted by stating the smallest number of perceptible steps 
between them (color threshold). This method applied to the partial 
errors in dominant wavelength, purity, and brightness, gives a good 
idea of the quality of color-films. The combination of the three indi- 
vidual errors into one final color error does not appear to be possible 
by any simple means. Information has been obtained from many 
investigations that greatest emphasis should be laid upon the correct 



May, 1937] 



BERTHON-SIEMENS COLOR PROCESS 



449 



dominant wavelength reproduction. Distortions in color purity are 
less disturbing and deviations in the relative reproduction of bright- 
ness are noticeable only in extreme cases. 

As an example of color distortion we studied the error in reproduc- 
tion of a simple continuous spectral region, for constant energy at 
the individual wavelengths. The spectral width of the reflectivity 



v; - 


S0p 





; 

r-v 
























^ 














+ 


6 - 




/ 










*^ 


-^ 




^ 




/ 
















1 


~** 


t '>^ 














































































































450 


% 

V 

c 

~. 

1 


sc 

7 


V 


ff0 

FIG. 3. 


600 


00 300 W COO 7t 

FIG. 1. 


: 




























r\ 
































A 




6 
































\ 


V 










' e. " 


--- 











X 












/ 


V, 


V 
































/ \ 


\ 






















L 
























) 








^ 




^v 






















\ 




10 \- 




























/ 


f 




I 




\\ 






















^ 


























SO 

t. 





I 


7 


- 


s 


'1 




V; 


^*50 500 550 600 
m; 

FIG. 2. 


< 0.3 

FIG. 4. 


06 0? ii 



FIG. 1. Optimal width of reflectivity band. 
FIG. 2. Color displacement of lenticular film in relation to threshold 

sensibility. 
FIG. 3. Saturation displacement of lenticular film in relation to threshold 

sensibility. 
FIG. 4. Spectral brightness curves in (1) taking, and (2} reproduction. 

function amounts to 50 m/i (Fig. 1). Dominant wavelength and 
purity errors are represented in Figs. 2 and 3, respectively, by the 
number of the perceptible color steps between the print and the origi- 
nal subject. The brightness distortions are exhibited in Fig. 4, in 
which curve 1 shows the visual brightness of 50-mAi spectral bands 
located in various parts of the spectrum, and curve 2 shows the 
brightness of the photographic reproductions of such colors. The 
color distortion represented here was computed on the assumption 



450 E. GRETENER [J. S. M. P. E. 

of ideal photographic rendering of the color components and upon 
the basis of conditions that could actually be realized in practice, 
for taking and projection. 

(2) Three-Color Additive Processes 

The colors used for the reproduction must be so chosen that the 
range of colors that can be represented comprises, as far as possible, 
all the colors occurring in nature. The use of pure spectral color is 
eliminated, since the screen brightness attainable with such sources 
is far too low for practical purposes. The combination of a white 
light-source with suitable red, green, and blue filters is the only type 
of source that can be considered for the projection of color reproduc- 
tions. The relative intensities of the projection colors chosen for use 
can be standardized by determining the relative amounts of these colors 
that must be projected simultaneously upon a white screen in order to 
produce a white image. The intensities of the projection components 
present in the synthesis of white are taken as units of the components, 
red, green, and blue in the colors of objects. The components of the 
color of an object are determined by producing a mixture of the par- 
ticular red, green, and blue lights to be used in projection such that 
the color of the mixture is identical with the color of the object. Such 
a system of standardized reproduction filters is shown in Fig. 5. 
This kind of component determination is not usable in practice; 
in its place the components are controlled by color filters. The action 
of such filters is determined by the spectral distribution of the light 
reflected from the colored objects. 

The colors of objects can be conceived as the superposition of spec- 
tral colors. The wavelength mixture radiated from an object is 
determined by the quality of the illumination and the reflectivity 
of the surface observed. The desired spectral transparency for the red, 
green, and blue filters depends upon the requirement of the highest 
possible similarity between the colored image and the original ob- 
ject. Thus, in the illustration of an object a picture is formed through 
a red filter whose energy distribution matches, point for point, the 
distribution of the red component in the original object. In the 
same way the green and blue components are formed. This energy 
distribution can be registered photographically. 

It is the function of the negative emulsion to rearrange the energy 
distribution of the parts of the picture into a proportional trans- 
parency distribution. This rearrangement must be independent of 






May, 1937] 



BERTHON-SIEMENS COLOR PROCESS 



451 



the spectral composition of the energy; and, over the greatest pos- 
sible range, must be independent of the absolute intensity of the color 
components. The first requirement must be fulfilled in order that, 
on the whole, simultaneous rearrangement of the energy into trans- 
parency values results for all colors. The second requirement repre- 
sents a condition of linearity, which makes the color independent of 
the brightness. For motion pictures the second condition is particu- 



SCH 




I/*, t. 




FIG. 5. 



FIG. 7. 




FIG. 6. FIG. 8. 

FIG. 5. Additive color mixture. 
FIG. 6. Distribution of light-rays in lens. 

FIG. 7. Path of light-rays through section of lenticular film. 
FIG. 8. Lens diagram showing circle of confusion. 

larly important, since, otherwise, a color would change with alternat- 
ing degrees of brightness. 

Now, if the three partial pictures are projected in register by 
means of the fundamental colors, then a colored picture of the original 
object is formed upon a white screen. 

THE LENTICULAR FILM 

(1) Fundamentals 

The method of three-color separation pictures is not usable in 
practice since, in addition to the great inconvenience of the method, 



452 E. GRETENER [J.S. M. P. E. 

registration of the three partial images in projection is not attain- 
able. On that account, the idea of combining the three color-separa- 
tion pictures to form a single picture has been tried. This method, 
for example, is used in the generally known Lumi&re autochrome 
plates. The filters are comprised of microscopically small elements 
in intimate contact with the emulsion. It has been attempted to 
introduce the color-screen method in films as well as in plates. Many 
difficulties arise in printing, since registration of the original screen 
and the printing screen is not attainable. Furthermore, projection 
of a color-screen film shows a "boiling" of the color-screen elements, 
produced by their differing distribution from picture to picture. 

The lens-screen film separates the color filter and photographic 
record. The three-part taking filter is placed in the pupil of the cam- 
era lens so that three component pictures are formed lying one over 
the other (Fig. 6). On account of the coordination of the directions 
of the image-forming bundles with the zones of the color filters, sepa- 
rate registration of the three component pictures is possible when a 
film provided with microscopically small cylindrical lenses on the 
celluloid side is used. 

The cylindrical lenses run parallel to the zones of the taking filters 
so that the energy values of the light-beams are recorded propor- 
tionally to the color components separated by the red, green, and blue 
zones of the taking filters. The whole picture is split into narrow 
strips by the lens screen (Fig. 7). The color values of the object, in 
recording, are distributed over the width of the strips. Therefore, 
the photographic emulsion contains, packed within it, the red, green, 
and blue components of the object. 

Reversal of the path of the rays during projection, in turn, co- 
ordinates the registration of the components in relation to the zones 
of the reproducing filters over the lens-screen, and thereby produces 
a naturally colored image upon a white projection screen. The 
superiority of the lenticular film in relation to the color-screen method 
follows from the separation of the filters from the photographic rec- 
ord. The print is possible fundamentally without screen superposi- 
tion. The disturbance caused by "boiling" (moire) of the screen does 
not occur upon projection, since each film lens is illuminated with the 
color mixture, and not in the three fundamental color strips. White 
would be represented upon the color-screen film by three strips of 
red, green, and blue, lying side by side; while in lenticular films, 
each individual lens appears homogeneously white. 



May, 1937] BERTHON -SIEMENS COLOR PROCESS 453 

Separation of the color-filter from the photographic recording, 
on the other hand, gives rise to the fundamental difficulties of lenti- 
cular film which now have been overcome. Falsities in color rendi- 
tion occur if the coordination between the recording strip and the 
filter zones of the objective is not exact. Correct color reproduction 
is possible only when the illumination systems for exposure, printing, 
and projection are free from vignetting; that is, the three-zone filter 
must be visible over its entire extent on every point of the image 
field. If this condition is not fulfilled, falsification of the color occurs, 
depending upon the position of the point observed in the image field. 
There are special requirements for the illuminating beam in printing 
and projecting. It is required that the illumination of the individual 
components be homogeneous over the entire image field. If the dis- 
tribution of energy at different elementary angles of the bundles of 
rays is not constant over the entire image field, then errors in color 
occur over the entire field on account of the splitting up of the image - 
forming rays into the three partial light-beams, red, green, and blue. 
As a further consequence, the separation of the filter from the photo- 
graphic emulsion makes it necessary to use a very accurate film-gate 
track. Any wrong orientation of the elementary lenses through 
curvature of the film leads to distorted color rendition. 

(2) Technology of Lenticular Films 

(a) Exposure Optics. In practice, objectives of different focal 
lengths are used for the exposure. Since projection must be consid- 
ered, the distance of the filter image from the film (during exposure) 
must be the same for all focal lengths. By using field lenses in the 
image plane, it is possible to make the distance of the filter image 
from the film identical for all focal lengths. This solution is of little 
practical importance, since the field lenses of the required power cause 
considerable loss of sharpness of the image when used in conjunction 
with the common objectives on the market. The requirement of 
freedom from vignetting and of uniform filter position for the expo- 
sure lens system limits the types of objectives suitable for lenticular 
film. The greatest difficulties arise at short focal lengths, so that 
wide-angle objectives are not yet available for color-film. The subdi- 
vision of the lens aperture into the three color zones leads to poor 
registration of the three separation images outside the focal plane. 
The cir.de of confusion of an illuminated point demonstrates the 
separation of the filter areas (Fig. 8). Therefore, color fringes result 






454 



E. GRETENER 



[J. S. M. P. E. 



on objects that are not sharply focused that are especially disturbing 
in close-ups where the requirements of depth of focus are extreme. 
For this type of scene, a special lens system has been developed. By 
placing a light-mixer in front of the objective, the pupil of the com- 
pound objective becomes identical for all three separation images. 
The circle of confusion of an illuminated point outside the sharp focus 
is in this way homogeneous as to color, and hence the appearance of 
color fringes is fundamentally eliminated. For clarity, in Fig. 9 only 




/lu 




FIG. 10. 



42 M 



FIG. 12. 




FIG. 9. Light-mixer in front of camera lens. 
FIG. 10. Diagram showing exposure variation from element to element 

of lens-screen. 

FIG. 11. Monochrome exposure behind several lens elements. 
FIG. 12. Alteration of step exposure through diffusion in the emulsion. 

the splitting of the light passing through area / is shown. In reality, 
the light passing through zones // and /// is also utilized. 

Looking from the film side, one sees the normal three-zone filter; 
while from the object side, a homogeneous pupil appears through the 
effect of the light-mixer. 

(b) The Photographic Recording of the Color Components. The 
lenticular film requires highest efficiency in the photographic emul- 
sions in regard to resolution. The amount of exposure given to the 
individual color zone images jumps discontinuously from zone to zone 
(Fig. 10). A step-like exposure is formed transversely to the cylin- 
drical lenses. Keeping in mind that the width of a zone image is of 
the same order of magnitude as the thickness of the silver bromide 



May, 1937] BERTHON-SlEMENS COLOR PROCESS 455 

layer, and considering that the silver bromide gelatin emulsion is a 
turbid medium having very strong light-diffusing properties, it is 
easy to recognize how difficult is the problem of recording the indi- 
vidual color components clearly. A clear view of the above-mentioned 
factors is obtained if exposures are made over a single color zone, 
excluding the other two zones (Fig. 11). It can be seen that due to 
the silver bromide diffusion, part of each color component reaches 
the areas provided for the other components in the photographic layer. 
The reproduction of a desired color takes place by superposition of 
the single exposures made through the red, green, and blue filters. 
Fig. 12 shows the alteration of the impressed exposure by silver bro- 
mide diffusion. 

Of special interest is the effect of the silver bromide diffusion in a 
non-stationary case ; that is, in the parts of the image where the color 
changes. The transparency function of a single zone is obtained by 
superposition of two exposure steps of opposite sign with their 
phases shifted by the width of a filter image (Fig. 13). Any desired 
color shade can be obtained through superposition by means of the 
transparency function of the single zones, and the color consequences 
can be viewed in the color triangle. 

In order to limit the effect of the silver bromide diffusion to a 
practical degree, Perutz, in connection with the Siemens Company, 
developed new types of emulsions. With proper adjustment of 
the taking and reproduction filters, and by maintaining certain sym- 
metrical conditions for the exposure distribution of the successive 
screen lenses, one can avoid the effect of the silver bromide diffusion 
upon the reproduced hue. The reduced saturation that results can 
be compensated by application of silver bromide solvents to the first 
developing solution. The effect is the same as if another gamma were 
adopted for the finest details than for the distribution of the image 
contrast as a whole. The concentration of the silver bromide solvents 
must be chosen in accordance with the properties of the photographic 
emulsion. 

In printing, another loss in saturation of the colors takes place due 
to the silver bromide diffusion. In this case, its effect can be checked 
by using relatively insensitive emulsions of very high resolving power. 

(c) The Printing Problem. In the printing process a correct- 
phase transfer of the black-and-white component records of the 
original relative to the edges of lenses of the film used for prints must 
take place. This task can be solved fundamentally without registra- 



456 



E. GRETENER 



[J. S. M. P. E. 



tion of the screen of the copies with that of the original. Fig. 14 
shows the course of the rays in contact printing. This procedure 
leads to moire* appearances by the action of both sets of lines in the 
screen, and this can be suppressed only with a very high-illumination 
aperture, which causes considerable loss of sharpness. The record 
lying behind an original lens is divided between several lenses of the 
positive film, which can also cause color fringing. 




FIG. 13. 



FIG. 15. 




TX 

- 



FIG. 14. 



FIG. 16. 



FIG. 13. Color changes occurring from one element to another. 

FIG. 14. Course of rays in contact printing. 

FIG. 15. Diagram of optical printing or projection. 

FIG. 16. Printing with multiple lens system. 



Another possibility of duplication of lenticular film consists in opti- 
cally printing the original upon the film used for prints. Fig. 15 shows 
the fundamental arrangement. In practice, this method involves 
great difficulties, since a printing lens of the required high aperture 
can not meet the requirements of image sharpness. With an aperture 
of the screen lenses of //2.5, an aperture of the printing objective of 
//1. 25 becomes necessary. Such an objective shows marked distor- 
tion of the images projected over the partial zones, and this leads 



May. 1937] BERTHON-SlEMENS COLOR PROCESS 457 

to very disturbing color fringing. Tests with modern high -aperture 
objectives have proved the hopelessness of this printing method. 
Fundamentally different ways had to be found. 

Three separate objectives of small aperture, which must present 
a large corrected-image field, are arranged in relation to the original 
film so that they individually register and record the red, green, and 
blue parts of the original. Fig. 16 shows the principal arrangement of 
the optics of the printer. The three-part images are registered through 
a mirror system. The printing optical system projects an image of 
the screen pattern of the original film, whereby the brightness of the 
individual lenses corresponds to the average brightness of the three 
components belonging to the lens in question. The screen of the film 
used for prints is brought into the focal plane of the original screen. 
Since the images corresponding to the three 
fundamental colors are arranged directionally 
with the three printing objectives, separate 
registration is possible of the part images over 
the lens screen, as in the original exposure. In 
this printing method, the moir6 appearances 
can be suppressed by using illumination of 
high aperture at right angles to the cylindrical 
lenses of the original film, without causing loss of 
sharpness. 

(d) Projection. The light requirements of the FIG. 17. Vari- 

lenticular film are about ten times as high as Density 11 h/'pro- 
those of the black-and-white film. The losses jector aperture, 
occur in the reproduction filters and through 
the reduction of aperture necessary to overcome vignetting of the 
illumination and reproduction systems. Besides the necessary in- 
crease in intensity of the projection light, special requirements are 
created as to the quality of the illumination in the projector aperture. 
The illumination systems used at present for black-and-white film 
all produce light rays with energy distribution that varies for different 
angles of the image field. The only necessary condition for black- 
and-white projection is an approximate proportionality of energy in 
the light rays belonging to the individual image points (Fig. 17). 
This alone fulfills the condition of approximately constant illumina- 
tion of the whole image field. 

The special conditions for the illumination of the lenticular film 
have been fulfilled by creation of a new type of arc lamp. Consider- 




Middle 



End 



458 E. GRETENER [J. S. M. P. E. 

able increase in the efficiency of illumination was attained as compared 
to present illumination systems. 

For small theaters a pure carbon arc was developed. In accordance 
with the form of the picture aperture, square carbons are used. A 
stabilizing arrangement at the hot end of the positive carbon effects 
the concentration of the total discharge upon the front surface, and 
protects the shell from oxidation. A magnetic field of a special type, 
with its axis parallel to the carbon axis, takes care of the stabilization 
of the arc. It is constructed so that rotation of the total discharge 
takes place at such high frequency that the homogeneity of the crater 







FIG. 18. Discharge form of new pure carbon 
lamp. 

for the illumination time of a single frame is assured during projec- 
tion. 

The space stability of the crater is so great that only a small safety 
margin of the crater image over the film-gate area is required. Fig. 
18 shows the burning form of the new pure carbon lamp. The positive 
and negative carbons have a common horizontal axis. The thermal 
lifting force on the arc is practically equalized through the electro- 
dynamic forces of the stabilizing field. Fig. 19 shows the stationary 
form of the positive carbon. 

Large theaters use high-efficiency lamps of very high intensity. 
Here also, new ways had to be found. In the existing high- intensity 
lamps, employing the Beck-effect, a deeply burned-out crater pre- 
vents the reflection of the luminous gas ball to the sides on account of 
the high crater walls. A considerable part of the current goes to the 
carbon shell, and is therefore lost for light production. In Fig. 20, 



May, 1937] 



BERTHON-SIEMENS COLOR PROCESS 



459 




FIG. 19. New positive carbon. 



the burning form and the cross-section through the crater of the posi- 
tive carbon of a commercial high-efficiency lamp are shown. The 
light-intensity distribution is not homogeneous, and is disturbed by 
flames emerging from the space between the shell and the core. 

In the newly developed high- 
efficiency lamp, square carbons 
are used. The lamp burns with 
absolute freedom from soot. 
Fig. 21 shows the discharge form 
of the new lamp. An intensely 
luminous gas ball is visible which 
extends well in front of the posi- 
tive carbon and therefore can 
radiate laterally without obstruc- 
tion. Fig. 22 shows the sta- 
tionary form of the positive car- 
bon. The flame gases are absorbed by a cover surrounding the 
negative carbon and escape through the bore of the reflecting mirror. 
In order to give an idea of the efficiency of the new lamp, it may be 
stated that with a current of 60 amperes, the density (intrinsic 



FIG. 20. Burning form and cross- 
section of commercial high-efficiency 
carbon. 




FIG. 21. Discharge form of new lamp. 




FIG. 22. Stationary form of new positive carbon. 

brilliance) in the crater was measured as 800 candles per sq. mm. 
Further increase of screen brightness above the limit reached with 
the new lamp could be attained by development of a special projec- 
tion screen. The screens that are at present in common use in motion 



460 



E. GRETENER 



[J. S. M. P. E. 



picture theaters, and which provide diffuse reflection, throw a great 
deal of the light from the projector upon the ceiling and the walls of 
the theater. The new reflector type of screen reflects the light only 
in those directions in which it is intended to go. Metal sheets are 
used in constructing the reflector, and small concave mirrors are rolled 
into them of such form that the desired diffusion diagram is obtained . 
The dimensions of the elementary mirrors must be kept very small, 
so that one million elementary mirrors are present in one square meter. 
Fig. 23 shows the increase achieved by concentration of the re- 
flected light, compared to the diffusing 
screen. On the basis of present conditions 
in German motion picture theaters, an in- 
crease of screen brightness by a factor of 3 
can be obtained through introduction of the 
new screen. The precision requirements of 
the elementary mirrors and the uniformity 
of impregnation must be extremely high, 
since otherwise changes of brightness ap- 
pear at the borders of the individual metal 
sheets. 

COLOR CONTROL 

Upon projecting a color-film, the viewing 
conditions are different from those when 
the object is viewed directly. The colored 
image appears luminous in a black frame. 
The connection with the surroundings is 
missing. The change from one scene to 
another occurs in jumps, so that the eye has no time to adapt itself 
to the changed mood. If pictures taken under varying conditions 
of illumination are combined into a single film, one gains the im- 
pression that, for example, evening scenes appear too red compared 
to noon exposures. In general, the eye adapts itself to the pre- 
dominating mood of the picture, so that in a change of scenery the 
next image at first shows a mood corresponding to the complemen- 
tary color of the preceding image. These difficulties can be avoided 
if, in changing an image, the mood of the following image is adapted 
to that of the preceding image. In an optical print from the len- 
ticular film, the light-rays belonging to the red, green, and blue 
component images can be controlled separately. By introduction 




FIG. 23. Increased 
screen brightness of new 
screen, compared with old 
screen. 



May, 



BERTHON-SIEMENS COLOR PROCESS 



461 



of diaphragms in the printing optics, the mood of a picture can be 
changed at will. 

In practice color control is effected in the following manner: At 
first, an uncontrolled copy is made, which is projected through an 
objective having controllable apertures in the red, green, and blue 
zones. In order to facilitate fine color tone changes, a controller is 
used which influences the objective diaphragms by movement of a 
single control member in the control triangle which is coupled to the 
diaphragm. The transmission of the movements of the control 
member on the filter zone diaphragms takes place electrically. The 
connection of the open filter areas to the positions of the control 
member in the control triangle is so arranged that changes of the 




R B 

Color Triangle 




Filter 



F B - r - g 



FIG. 24. Diagram showing method of color ratio 
control. 



white point of the projection correspond to the movement of the con- 
trol member in the control triangle. If the control member is ad- 
justed for the white point of the control triangle, the projected image 
reproduces a mood as it actually was at the time of the exposure. If 
one describes a circle about the white point with the control member, 
the color tone runs through the entire color circle around the white. 
Fig. 24 shows schematically the construction of the projector control. 
A trained judge of color quality decides the necessary correction 
for the individual images with the aid of the control triangle. In 
order to judge the effect of the correction of the individual scenes 
during the run of a film, a perforated strip is made which automati- 
cally controls the diaphragms of the projector objective (on a contact 
apparatus). The start of the adjustments of the diaphragms is ef- 
fected by a film contact. After the necessary fine adjustments, fur- 
ther prints are made by control of the printer diaphragms by means of 



462 E. GRETENER [j. s. M. P. E. 

this controlling strip. The threater prints are then uniform in the 
color tone assigned to them by the color-quality judge on the control 
projector, without the use of any control arrangement upon the 
theater projectors. 

The physical principles upon which the lens-screen process is 
based have been known since the close of the last century. 1 The 
method concerned the combination of three color-separations into a 
single picture with the aid of a three-zone filter within an objective. 
At that time, for separate registration of the three color images, a 
line-screen diaphragm was placed in front of the photographic layer. 
In 1908, the Frenchman, Berthon, used, instead of the impracticable 
line-screen, the lenticular screen in the form of microscopically small 
cylindrical lenses. 2 The lenses are impregnated by rollers into the 
celluloid support. Berthon is then the inventor of the embossed film. 
The efforts of the French group, which had set itself the task to de- 
velop to a commercial stage a color-film process based upon the em- 
bossed film, appeared to have led to a certain success by 1930. It 
was believed that a printing process for lenticular film had been de- 
veloped that permitted printing the original in theater quality. At 
that time further development was taken over by Siemens, and in- 
vestigations in the laboratory very soon showed that the process 
was still very far from the commercial stage. A considerable number 
of technical difficulties arose which have required about six years of 
intensive laboratory studies with great expenditures. 

CONCLUSION 

The general principle of the additive color film is explained: The 
composition of the colors in reproduction is discussed and analysis of 
the components is described. Absolutely natural reproduction is 
practically impossible; the task is to make the deviations as small 
as possible. 

The lenticular film differs from the color-screen film by separation 
of screen and filter. This has the following advantages : more liberal 
choice of dyes, easier and cheaper manufacture of the film, no 
"boiling" in the reproduction. The disadvantages could be removed 
by proper selection and design of new taking optics, as well as by 
improved film-gates. The moire-free printing is done with a special 
printing lens and large illumination aperture. 

The photographic difficulties, which were eliminated in the films 
developed by Perutz, are described. 



May, 1937] BERTHON-SiEMENS COLOR PROCESS 463 

The necessary projection brilliance is attained by new arc lamps 
and grooved aluminum reflectors for screens. In printing, the color 
mood can be controlled in an easily predetermined manner. 

REFERENCES 

1 AHRIMAN LIESEGANG, R. E.: "Von der Zukiinftigen Photographic ein 
neues Prinzip der Farbenphotographie," Phot. Arch., 37 (Aug., 1896), No. 96, 
p. 249. 

* French Pat. No. 399,762. German Pat. No. 223,236 (1910). 



RECENT DEVELOPMENTS IN MAGNETIC SOUND 
RECORDING* 

S. J. BEGUN** 



Summary. Although first work in magnetic sound recording occurred probably as 
far back as 1900, it has been only recently that any considerable interest has been 
aroused in it. The present paper describes the principles underlying the methods oj 
recording and reproducing by magnetic means, both by longitudinal and perpendicular 
magnetization, and a steel tape machine used in European broadcasting stations is 
illustrated. The advantages of the magnetic systems of recording over other systems 
are outlined. 



Looking backward, it seems rather strange that of the three known 
methods of recording sound, only the mechanical and photographic 
methods have come into wide use, while the magnetic method has 
been hardly touched. 

Magnetic sound recording was invented by the Danish physicist, 
Valdemar Poulsen, in 1900. Quite a number of patents were granted 
to Poulsen and his associates on various recording methods and 
machines devised by them in the early part of the century. Poulsen 
has shown that when a carrier made of magnetic material, such as a 
steel wire or tape, is moved between magnetic poles magnetized 
by speech currents, the successive elements of the carrier are con- 
verted into a succession of elemental magnets of varying degrees 
of magnetization. By passing a carrier with such elemental magnets 
between a similar set of reproducing poles, corresponding speech 
currents are induced in the coils surrounding the poles. 

Although various companies in Europe, and the American Tele- 
graphone Company in the United States, took up Poulsen's inven- 
tion, little progress was made in the commercial exploitation of 
magnetic recording and its practical application was abandoned. 
However, the inherent advantages of magnetic recording have 
again aroused interest in its possibilities, and magnetic recording 
machines have within recent years appeared on the market in 

* Received March 5, 1937. 
** Guided Radio Corp., New York, N. Y. 
464 



MAGNETIC SOUND RECORDING 



465 




Europe, particularly in Germany. It is safe to predict that before 
long magnetic sound recording machines will become of commercial 
importance in the United States. 
A brief description of the fun- 
damental principles underlying 
magnetic recording may be of 
interest in this connection. 
Poulsen has already shown that 
magnetic recording may be ac- 
complished either by aligning 
two pole-pieces upon the oppo- 
site sides of the magnetic carrier 
to produce a succession of per- 
pendicularly magnetized elemen- 
tal magnets perpendicular to the 
direction of the travel of the 
carrier, or by offsetting the re- 
cording pole-pieces on opposite 
sides of the carrier to produce a 
succession of longitudinally magnetized elemental magnets in the 
direction of the travel of the carrier. 

Sound recording and reproduc- 
ing by either of these methods 
of magnetization involves three 
operations, namely, the blotting, 
the recording, and the reproduc- 
ing operation. Each of these 
operations is carried out with 
substantially the same kind of 
magnetic structure, which, de- 
pending upon its function, is 
designated as the blotting or 
obliterating head, the recording 
^ head, and the reproducing head. 



l * P 



FIG. 1. Magnetic head arranged for 
perpendicular magnetization. 




FIG. 2. Magnetic head arranged for 
longitudinal magnetization. 



Fig. 1 illustrates an arrange- 
ment of a magnetic head for per- 
pendicular magnetization. The 
magnetic pole-pieces P are aligned upon opposite sides of the 
magnetic carrier, for instance, a flat steel tape of thickness T 
moving in the direction of the arrow. Current flowing through a 



466 



S. J. BEGUN 



[J. S. M. P. E. 



magnetizing winding W surrounding the pole-pieces P produces a 
main magnetic flux M, which passes perpendicularly through the 
tape element lying between the pole-pieces, and a leakage flux L 
which spreads out in the tape on both sides of the zone lying 
between the pole-pieces P. 

Fig. 2 illustrates an arrangement of a magnetic head for longitudi- 
nal magnetization. The two pole-pieces P, upon opposite sides of 
the carrier, are offset by a distance Z, and the winding W surrounding 
the pole-pieces produces a main flux M which passes from one pole 
piece through the tape to the other pole-piece, and a leakage flux 




FIG. 3. Hysteresis loop of steel carrier. 

L extending through the tape in a direction opposite to the main flux. 

Although there are certain practical differences between the re- 
cording by the perpendicular and longitudinal methods in the actual 
recording, the magnetic phenomena underlying the blotting, record- 
ing, and reproducing operations are similar for both methods, and 
may be analyzed by reference to the magnetization curve of the 
magnetic carrier, shown in Fig. 3, on the theoretical assumption 
that the magnetizing effect of the pole-pieces is confined to a point 
of the moving tape. 

After the tape material has been initially magnetized, following 
the broken curve from zero to the saturation point 5, the magnetiza- 
tion curve takes the form of the well known hysteresis loop, and a 



May, 1937] 



MAGNETIC SOUND RECORDING 



467 



tape element may at any time be magnetized to its saturation point 
5 by a magnetizing current 1. 

In the recording process, the tape travels at a constant speed, and 
each element of the tape passes between a pair of obliterating mag- 
nets which are magnetized to saturation by the current /. This 
forces through each tape element passing between the magnets a 
uniform saturating flux corresponding to a point 5 on the hysteresis 
loop. This saturating flux blots out all previous magnetizations 
of the tape, and all previous recordings are accordingly wiped out. 
As the element of the tape, thus saturated, moves away from the 
blotting magnets, the magnetization decreases, following the de- 




FIG. 4. Magnetic sound recorder (Electrical Communication; July, 1936). 

scending branch of the magnetization curve to the value of the 
remanent magnetization R, which remains uniform for the successive 
elements of the tape passing through the blotting magnet. 

Continuing the motion of the tape, each uniformly magnetized 
element passes between a set of recording magnets, which are sub- 
jected to the joint action of a demagnetizing direct current I and 
an alternating speech current I s , which force a resultant demagne- 
tizing flux through each element and reduce the magnetization along 
the descending branch of the magnetizing curve to a point below 
R for example, to the point B m so that upon leaving the demag- 
netizing zone of the recording magnets, the magnetization of the 
tape element increases along the minor ascending loop from B m to 
B m ' and attains a new value of remanent magnetization, B m '. 



468 



S. J. BEGUN 



[J. S. M. p. E. 



Under the action of the oscillatory recording current I s , the suc- 
cessive elements of the uniformly magnetized tape will be subjected 
to successive demagnetizations oscillating between the values B m 
and B n , along the descending branch of the curve, so that upon 
leaving the recording magnets, the magnetizations of the successive 
tape elements will increase along minor ascending loops to corre- 
sponding remanent magnetizations varying between B m ' and B n ', 
converting the recorded carrier into a succession of elemental mag- 
nets which form along the carrier a magnetic wave B r , corresponding 
to the recorded sound. 



6 

2 

10 1 
8 
6 

It 

2 

\ 


























^-- 


, 
































g 






S 


























, 


X 










\ 




















i 


/ 
















\ 
















/ 


S 
































A 


/ 


































/ 


























































































































































































































































































































6 8 JO 2 2 168 


JO 3 2 I 6 8 


1C 



FIG. 5. Unequalized response curve of commercial magnetic 
recorder. 

In the reproducing process, the recorded carrier is passed with 
the same speed and in the same direction between a set of similar 
reproducing poles, and the remanent magnetomotive forces of the 
elemental magnets forming the recorded magnetic sound-wave force 
through the reproducing cores magnetic fluxes that induce in the 
surrounding coils alternating currents for reproducing the recorded 
sound. In distinction from photographic sound recording systems, 
the induced reproducing currents are not proportional to the am- 
plitude of the recorded sound-waves, but to the rate of change of 
the magnetic flux waves along the carrier. 

In the practical design of magnetic recording systems, the in- 



May, 1937] MAGNETIC SOUND RECORDING 469 

ducing zone of the magnet poles can not be confined to a single 
point of the wire or tape, and the magnetic phenomena of the re- 
cording and reproducing processes are much more complicated. In 
order to approach such ideal operating conditions, the thickness 
of the pole-faces and the inducing zones of the carrier are made as 
small as practicable. This is also important, because in order to 
accomplish satisfactory reproduction, the length of the induced 
zone must be smaller than one-half the wavelength of the sound-waves 
recorded on the carrier. 

Furthermore, while under the theoretical conditions described 
above, maximum efficiency would require a large demagnetizing 
direct current and large recording currents to permit utilizing the 
entire straight portion of the descending branch of the magnetization 
curve for recording practical limitations make this undesirable. 

When using perpendicular magnetization, as shown in Fig. 1, the 
recording flux entering the tape spreads out laterally beyond the 
zone engaged by the poles, and the increased breadth Z of the in- 
duced carrier zone depends not only upon the thickness of the poles 
P, but also upon the thickness T of the tape. As a result, each 
recorded tape element leaving the poles is subjected to an additional 
magnetizing action by the spreading stray field of another recording 
flux, which distorts the recording effected directly beneath the poles. 

In order to make satisfactory records by perpendicular magne- 
tization, a very thin tape must be used so as to reduce to a minimum 
the lateral spread of the recording flux. Such recording permits 
operation with a low tape speed, but only thin weak steel tape can 
be used as a recording medium and the perpendicularly magnetized 
elements of the thin tape are able to store only little energy. 

With longitudinal magnetization, as shown in Fig. 2, the limi- 
tations with regard to the thickness of the magnetic carrier are 
avoided, and either a strong steel wire or a strong steel tape may be 
used as the recording medium. The thickness of the pole-pieces 
P and the slot distance Z by which the pole-pieces are offset are so 
proportioned as to assure that the recording flux magnetizes longi- 
tudinally only a very small length of the carrier, while keeping the 
slot distance Z a minimum. 

In operating with longitudinal magnetization, the effect of the 
leakage flux must be taken into consideration. As shown in Fig. 2, 
each successive longitudinally magnetized carrier element leaving 
the zone of the magnets in the direction of the arrow, passes through 



470 S. J. BEGUN [J. S. M. p. E. 

a zone that is magnetized by the oppositely directed leakage flux L, 
and its magnetization will be reduced from the remanent value R, 
along the downward slope of the curve of Fig. 3 to RI, so that the 
carrier will leave the zone of the leakage field with the new remanent 
magnetization RI'. Accordingly, blotting magnets operating with 
excessive magnetizing currents impair the subsequent recording 
process, because the resulting leakage flux reduces the magnetiza- 
tion available for the recording to the straight portion of the de- 
scending curve below the point RI. By designing the blotting mag- 
nets for operation with a small leakage flux, this disturbing effect 
of the leakage flux upon the recording operation may be rendered 
negligible. 

Furthermore, the leakage flux of the recording magnets tends to 
reduce the magnetization of the recorded tape elements moving 
through the zone of the leakage flux. By designing the recording 
magnets to operate with a relatively small demagnetizing direct 
current, and with sound currents restricted to only a part of the 
straight portion of the descending magnetization curve below the 
point R r , Fig. 3, the disturbing effect of the leakage flux upon the 
recording action is rendered negligible. 

Although longitudinal magnetization requires a greater length 
of wire or tape for recording the same length of speech than per- 
pendicular magnetization, recording by longitudinal magnetization 
is much more practicable. This is due to the fact that all available 
magnetic sound carriers, such as steel wire or tape, are produced 
by drawing processes, and drawn steel material can be magnetized 
to a much greater extent in the longitudinal direction than in the 
transverse direction. Accordingly, the succession of longitudinally 
magnetized elements of a tape recorded by longitudinal magnetization 
in the manner here described supply much greater magnetomotive 
forces for inducing in the coils of the reproducing magnets large 
sound-reproducing currents without disturbing noises than could be 
obtained with perpendicular magnetization. 

Magnetic sound recording machines operating in accordance with 
these principles are now used in most of the European and Canadian 
broadcasting stations, and are also beginning to enter into the dic- 
tating machine field. 

A magnetic tape sound recording machine built under the direction 
of the author for the German and other European broadcasting 
stations by C. Lorenz Aktiengesellschaft of Berlin is shown in 



May, 1937] MAGNETIC SOUND RECORDING 471 

Fig. 4. This machine makes a continuous recording of .'M minutes. 
Fig. 5 shows an uncorrected frequency response curve of the repro- 
ducing head of this machine. By a suitable corrective network, this 
response curve can be made uniform between 100 and 8000 cps. 
A recording made with such a machine will remain intact for several 
years, and will be unaffected by temperature changes below 250 C. 
Summarizing, it may be stated that magnetic recording has the 
following advantages over other systems: It requires no processing, 
and the record may be reproduced immediately after it is made 
without impairing the recording. Each portion of the recorded 
sound may be repeated as many times as desired without affecting 
its quality. In distinction from mechanical and photographic re- 
cordings, the quality of a magnetic recording may be checked as 
soon as it is made. Since the recording medium is not subjected 
to mechanical or chemical changes, the same sound carrier may be 
used over and over again. Recorded words or sentences may be 
replaced by others without affecting the remainder of a completed 
recording. By combining the blotting magnets with the recording 
and reproducing magnets into a single head, each new recording 
operation wipes out the prior recording without loss of time and 
effort. The recording mechanism is not sensitive to external vi- 
brations, and may be used for making records while riding on a 
train, in an automobile, or in an aeroplane. The simple mecha- 
nism can be operated by a child. There is accordingly little doubt 
that the new magnetic recording devices are bound to find new 
fields of use in the United States. 

BIBLIOGRAPHY 

POULSEN, V.: U. S. Patents: 661,619, Original wire recording machine. 
873,083, Wire and tape recording with perpendicular magnetization by oblitera- 
ting magnets and polarized recording magnets. 789,336, Wire spool machine. 
873,078, Longitudinal magnetization recording. 

POULSEN, V.: "Das Telegraphon," Annal. der Physik.3 (1900), pp. 754-760. 

RELLSTAB, L. : "Der Telegraphon," Electrotechnische Zeitsch., 22 (1901 ), p. 57. 

WEST, T. H.: "Der Telegraphon," Electrotechnische Zeitsch. (1901), No. 22, 
p. 182. 

STILLE, K. : "Die Elektromagnetische Schallaufzeichnung," Electrotechnische 
Zeitsch. (1930), p. 449. 

BEGUN, S. J.: "Telephon-Aufnahme Maschinen zum Emfang von Schnell- 
Telegraphie," Telegraphen-Praxis (1931), No. 5. 

BEGUN, S. J.: "Die Diktiermaschine im Grossbetrieb," Electrotechnische 
Zeitsch. (1932), p. 204. 



472 S. J. BEGUN 

HORMANN, E.: "Zur Theorie der Magnetischen Tonaufzeichnung," Electro- 
technische Zeitsch., 9 (1932), p. 388. 

MEYER, E., AND SCHULLER, E.: "Magnetische Schallaufzeichnung auf Stahl- 
bander," Zeits.fur Techn. Physik, 13 (1932), p. 593. 

BEGUN, S. J. : "Beitrag zur Theorie der Elektromagnetischen Tonaufzeichnung 
auf Stahldraht," Dissertation Sonderdruck Studentenhaus Chariot tenburg, 1933. 

HICKMAN, C. N.: "Delayed Speech," Bell Lab. Record, 11 (June, 1933), No. 6, 
p. 308. 

BRAUNMUHL, H. J. VON: "Magnetische Schallaufzeichnung im Rundfunkbe- 
trieb," Funk (Dec., 1934), No. 40, p. 483. 

FREDERICK, H. A.: "Recording and Reproducing Sound," Rev. Sci. Instr., 5 
(May, 1934), No. 5, p. 177. 

"Magnetic Recording and Reproducing," Wireless World, 34 (Jan., 1934), 
No. 1, p. 8. 

RUST, N. M.: "Marconi-Stille Recording and Reproducing Equipment," 
Marconi Review (Jan., 1934), No. 1, p. 1. 

HANSEN, W. H.: "Das Magnetophon," Electrotechnische Zeitsch., 56 (Nov., 
1935), p. 1232. 

MALUNA, R. F. : "A Mirror for the Voice," Bell Lab. Record, 13 (March, 1935), 
No. 7, p. 200. 

SCHULLER, E.: "Magnetische Schallaufzeichnung," Electrotechnische Zeitsch., 
56 (Nov., 1935), p. 1219. 

VOLK, T.: "Magnetophon, das neue Tonaufzeichnungsgerat der A. E. G.," 
A. E.G. Mitteilungen (Sept., 1935). 

VOLK, T. : "Magnetophon, einneues Tonaufzeichnungsgerat," Filmtechnik, 11 
(Oct., 1935), p. 229. 

HAMILTON, H. E.: "The Blattnerphone Its Operation and Use," Electrical 
Digest (Dec., 1935), No. 12, p. 347. 

BEGUN, S. J.: "Dies neue Lorenz Stahlton-Bandmaschine," Lorenz Berichte 
(Jan., 1936), No. 1, p. 3. 

SCHRAGE, W. F.: "Sound Recording on Magnetic Materials," Radio Craft, 7 
(March, 1936), No. 3, p. 537. 

BEGUN, S. J.: "The New Steel Tone Tape Machine," Electrical Communica- 
tion," 15 (July, 1936), No. 1, p. 62. 



ICONOSCOPES! AND KINESCOPES^ IN TELEVISION* 
V. K. ZWORYKIN** 

Summary. A description is given of the theory and performance of the television 
system based upon the Iconoscope and Kinescope. The cathode-ray transmitting and 
receiving tubes used in the system are discussed in detail, considering not only the 
physical principles involved but also the operating characteristics. 

In conclusion, an account is presented of the television project at the Empire State 
Building (New York), describing the terminal equipment and giving examples of 
the 441-line picture that can be transmitted and received. 

Two extremely important elements in any television system are the 
pick-up device which converts the light-image into electrical signals, 
and the viewing arrangement transforming the electrical signals back 
into visible images. In fact, the success or failure of a television sys- 
tem depends perhaps more upon these two links than upon any other 
part of the chain. In its present project RCA Manufacturing Com- 
pany is using the Iconoscope and Kinescope for dissecting and resyn- 
thesizing images, and it is the purpose of this discussion to explain the 
operation of these instruments and point out the reasons for selecting 
them in preference to other devices designed to serve as pick-up and 
viewing equipment. 

Historically, the development of any form of television had to await 
a means of converting a light-signal into a corresponding electrical 
impulse. This step became possible through the discovery of the 
photoeonductive properties of selenium in 1873. Within two years 
after this discovery, Carey proposed to make use of the properties of 
selenium in the solution of the problem of television. His suggestion 
was to construct a mosaic consisting of a great number of selenium 
cells, in a sense imitating the retina of the human eye. These cells 
were to be connected to shutters or lamps in corresponding positions 
on a viewing board. Although the suggestion was made in 1875, the 
device was not put into operation until 1906, when Rignoux and 

*Reprinted from RCA Review, I (July, 1936), No. 1, p. 60. 
**RCA Manufacturing Co., Camden, N. J. 

t From the Greek Icon, meaning an image, and scope, signifying observation. 
J From the Greek Kineo, meaning movement. 

473 



474 V. K. ZWORYKIN [J. S. M. P. E. 

Fournier used this arrangement to transmit simple patterns and let- 
ters. Their mosaic consisted of a checker-board of sixty-four sele- 
nium cells. Each cell was connected to a shutter on a viewing screen 
which was also made up of sixty-four elements in positions corre- 
sponding to those in the pick-up screen. When a picture was pro- 
jected upon the selenium cells the resistance of those illuminated de- 
creased, allowing an electric current to flow which opened correspond- 
ing shutters on the viewing screen. A light behind thsese shutters 
made the reproduced picture visible. 

The idea of dividing the picture into elements, converting the 
illumination on each element into electric current, and sending the 
signal from each over individual wires is practicable for a small num- 
ber of divisions or picture elements and for transmission over short 
distances, but is useless as a means of producing pictures of the stand- 
ard required of television today. 

The next step was proposed by Nipkow in 1884. Instead of using 
individual wires connecting each picture element, he suggested send- 
ing the information from one element at a time over a single communi- 
cation channel and then reassembling this information again at the 
viewing screen. This process was to be carried out at such a rate that 
the picture appeared continuous due to persistence of vision. The 
means proposed to accomplish this point-by-point transmission was 
the scanning disk. At the time of its invention the necessary technic 
of handling and amplifying small currents had not yet been developed 
so that it was a number of years before this scanning principle could 
be put to practical use. However, the principle was sound, and the 
scanning principle has been the basis of all television systems since 
then. 

While this development represents a great step forward, it was 
attained only at considerable expense of available picture signal. The 
loss is due to the fact that each element contributes to the picture 
only for a small fraction of the total time, whereas with the first sys- 
tem suggested each element operated continuously. To make this 
clear, consider again the simple sixty-four element mosaic used by Rig- 
noux and Fournier. Each photoelectric element was connected to the 
viewing screen by a separate conductor and the picture to be trans- 
mitted was projected continuously upon all the elements, so that a 
signal current passed through every light-sensitive element all the 
time. To reduce the scanning system to a comparable case, assume 
that we have the same mosaic of sixty-four photosensitive elements, 



May, 1937] ICONOSCOPES AND KINESCOPES 475 

but that they are all connected to a common communication channel. 
The elements are covered with shutters (*. e., the scanning disk) which 
allow the light from only one element of the picture at a time to reach 
its corresponding photocell. These shutters are opened one at a time 
in rotation, covering the entire picture twenty or thirty times a sec- 
ond. Thus each light-sensitive element is operating only for a frac- 
tion of the total time, equal to one over the number of picture ele- 
ments, in this case one sixty-fourth of the time. 

In order to -regain this lost signal and yet retain the principle of 
scanning, the development of the Iconoscope was undertaken. To 
illustrate the method of attack, consider again the sixty-four element 
array of photocells. Instead of scanning the elements with shutters, 
assume that each element is connected to the contact points of a 
switch which connects them in rotation to the main communication 
channel. Thus the scanning is accomplished by means of a commu- 
tator switch. 

So far, we have gained nothing over the previous method of scan- 
ning, but now if a condenser is placed across each of the photocells 
in such a way as to accumulate the entire charge released by the ac- 
tion of the light during the time the element is not connected to the 
communication channel, this charge can be used when the commuta- 
tor switch again makes contact with this element. Therefore, photo- 
electric current is being released by every element continuously and 
the charge is stored in the condenser belonging to that element until 
it is needed at the end of a scanning cycle. 

The reduction of this principle to some practicable form is obviously 
a difficult problem. The number of individual photocells and con- 
densers required for a 441 -line picture with a 4 to 3 aspect ratio will 
be of the order of 260,000 units, and it is quite apparent that a screen 
composed of that many conventional photocells and condensers is out 
of the question. 

A solution devised by the author some years ago was to build up a 
mosaic screen which contained the equivalent of a vast number of 
photosensitive elements and condensers. This mosaic was mounted 
in a cathode-ray tube in such a way that an electron beam could be 
used to commutate the elements. Fig. 1 shows one of these tubes to- 
gether with its associated circuits, taken from one of the author's 
early patents. 1 Aside from the advantage gained through the appli- 
cation of the principle of storing the charge on each element for the 
entire picture time, this tube had the additional advantage that it 



476 



V. K. ZWORYKIN 



[J. S. M. P. E. 



involved no mechanical moving parts such as a scanning disk, mirror 
screw, or drum, the scanning being done electrically. 

Although the first of this type of tube was built as far back as 
1923, many years of research and development had to be undertaken 
before it was perfected sufficiently to meet the requirements of a sat- 
isfactory television system. The history of this development is in- 



rrriG; 




FIG. 1. Cathode-ray tube containing mosaic mounted so as to permit 
electron beam to commutate elements (taken from U. S. Patent No. 1,691,- 
324; V. K. Zworykin "Television System") 

teresting, but is somewhat outside the scope of this paper, which will 
be limited to a discussion of the tube as it is today. 

The Iconoscope, as this type of tube has been named, is shown 
photographed in Fig. 2. It consists of an electron gun and photo- 
sensitive mosaic enclosed in a highly evacuated glass envelope. The 
arrangement of these elements is shown diagrammatically in Fig. 3. 

The electron gun produces a narrow pencil of cathode rays which 
serves, as will be shown later, as a commutator to the tiny photocells 
on the mosaic. The gun is in reality a form of electron projector 
which concentrates the electrons from the cathode onto the mosaic in 



May, 1937] 



ICONOSCOPES AND KINESCOPES 



477 



a very small spot. The electron optical system consists of two elec- 
tron lenses formed by the cylindrically symmetrical electrostatic 
fields between the elements of the gun. Fig. 4 shows diagrammati- 
cally the arrangement of this gun, together with the equipotentials of 
the electrostatic fields making up the electron lenses. Below this 
diagram is the approximate optical analogue. Details of the gun 
construction are as follows : The cathode is indirectly heated with its 
emitting area at the tip of the cathode cylinder. It is mounted so 
that the emitting area is a few thousandths of an inch in front of an 
aperture in the control grid. A long cylinder with three defining aper- 
tures whose axis coincides with that of the cathode cylinder and con- 




FIG. 2. The Iconoscope. 

trol grid serves to give the electrons their initial acceleration, and is 
known as the first anode. A second cylinder, coaxial with the first 
anode and of somewhat greater diameter, serves as second anode and 
gives the electrons their final velocity. The second anode is in gen- 
eral formed by metalizing the neck of the Iconoscope bulb, as shown 
in Fig. 3. The gun used in the Iconoscope is designed so that it will 
concentrate a beam current of from one-half to one microampere into 
a spot about five mils in diameter. Under ordinary operating condi- 
tions, a potential of about a thousand volts is applied between the 
cathode and second anode, and the voltage of the first anode is ad- 
justed until minimum spot size is obtained. The exact value of the 
beam current to be used will, of course, depend upon the type of pic- 
ture to be transmitted and the exact conditions of operation. 



478 



V. K. ZWORYKIN 



[J. S. M. P. E. 



The beam from the gun is made to scan the mosaic in a series of 
parallel horizontal lines repeated at thirty cycles per second. This is 
accomplished by two sets of magnetic deflecting coils arranged in a 
suitable yoke and slipped over the neck of the Iconoscope. These 
sets of coils are driven by two special vacuum-tube generators supply- 
ing a saw-toothed current wave, one operating at picture frequency 
and supplying the vertical deflecting coils, the other at horizontal line 
frequency driving the second set of coils. 

The element that characterizes the Iconoscope is the mosaic. It 
consists of a vast number of photosensitive globules mounted upon a 




"71 


^ 




I 


pi 


P 


/a 




> 








X 




i. 




lilili 


HI|I|I 


V 

\ i 


=? 


W 



CO/I. 5 1 

FIG. 3. Arrangement of elements of Iconoscope. 



thin mica sheet in such a way that they are insulated from one an- 
other. The back of this sheet is coated with a conducting metallic 
film which serves as a signal plate and is connected to the input of the 
picture amplifier. The appearance of the mosaic is shown in Fig. 5. 
Such a mosaic may be formed in a variety of ways. For the standard 
type of mosaic the silver globules are formed by reducing particles of 
silver oxide dusted over the mica. Under the proper heat treatment 
the silver globules reduced from the oxide will not coalesce but will 
form individual droplets. These droplets are sensitized after the 
mosaic has been mounted in the tube and the tube evacuated. The 
sensitization is similar to that used in the ordinary cesium photocell, 
that is, the silver is oxidized, exposed to cesium vapor, and then heat- 



May, 1937] 



ICONOSCOPES AND KINESCOPES 



479 



treated. The spectral response depends upon the details of the acti- 
vation schedule and can be varied to meet the operating conditions 
required of the Iconoscope. The spectral characteristic of the present 
standard Iconoscope is shown in Fig. 6. It is evident that the Icono- 
scope with a quartz window is sensitive from well into the infrared, 
through the visible, and into the ultraviolet. Actual tests have pro- 
duced images using radiation from 2000 A down to more than 9000 A. 
The mica upon which the silver droplets are mounted serves to in- 




APPROX. OPTICAL ANALOGY 




INDICCS Of RE FRACTION : /*, ^ >/f, > 

FIG. 4. Diagrammatic arrangement of electron gun. 



sulate them from one another and, further, is made thin enough so 
that the capacity between each globule and the metallic signal plate 
will be reasonably large. The uniformity of cleavage sheets of mica, 
together with their excellent insulating properties, low dielectric hys- 
teresis, and low loss make them very suitable for this purpose. 
Other insulating materials, however, can be used ; for example, a thin 
film of vitreous enamel upon a metal signal plate has proved very sat- 
isfactory. 

The mosaic is mounted in the tube with the silver beads facing the 
beam. In order that the optical image may be focused on its front 



480 



V. K. ZWORYKIN 



[J. S. M. P. E. 



surface, it is placed in the tube in such a way that a normal to its face 
makes an angle of 30 degrees to the axis of the electron gun. 

In essence, the Iconoscope may be thought of as a plain mosaic 



f>H/OTOi:LEMNTS Of 







FIG. 5. Appearance of mosaic in Iconoscope. 

made up of a great number of individual photocells, all connected by 
capacity to the common signal plate and commutated by the scan- 
ning beam. The fundamental cycle of operation is as follows : 

Every silver globule making up the mosaic is photosensitized so 







\ 




















\ 


















\ 












w 
> 








V 










*> 
Si 

* 








\ 


^^ 




















* ^"""**'^^ l ^^ 


-. 




























fo 




60 


oo ' 


re 


M 


m 



WA VEL e*6 TH (AN&STKOM UNITS) 

FIG. 6. Spectral characteristic of mosaic. 

that when a light-image is projected upon the latter the light causes 
electrons of a number proportional to the light brilliance to be emitted 
from each illuminated minute photosensitive area. The resulting 



May, 1937] ICONOSCOPES AND KINESCOPES 481 

loss of electrons leaves each photosensitive area at a positive potential, 
without respect to its initial condition, which potential is then pro- 
portional to the number of electrons that have been released and con- 
ducted away, so that the mosaic tends to go positive at a rate propor- 
tional to the light falling upon it. As the electron beam scans the 
mosaic, it passes over each element in turn, releasing the charge it has 
acquired and driving it to equilibrium. Due to the fact that each ele- 
ment is coupled by capacity to the signal plate, the sudden change of 
charge of the elements will induce a change in charge on the signal 
plate and result in a current pulse in the signal lead connected to the 
amplifier. The magnitude of these pulses will be proportional to the 
intensity of the light falling upon the scanned element. Thus the 




FIG. 7. Equivalent circuit of single element of the mosaic, 
for storage system. 

signal output from the Iconoscope will consist of a chain of current 
pulses corresponding to the light distribution over the mosaic. This 
chain can be resynthesized at the receiver into a reproduction of the 
original image, as will be described later. 

To clarify this cycle the equivalent circuit representation of a 
single element is shown in Fig. 7. The beam is represented by the 
switch and series resistance R. This switch may be considered as 
being open except at such times as the beam is actually on the ele- 
ment. When the scanning beam moves off the element, the photo- 
emission from it starts to charge the condenser C, the rate of accumu- 
lating charge being proportional to the illumination of the element. 
In the next scanning cycle, the beam again sweeps over the element, 
closing the switch and discharging the element. During this dis- 
charging cycle the entire charge accumulated during the time the 
beam was not on the element must now flow through the input resistor 






482 



V. K. ZWORYKTN 



[J. S. M. P. E. 



R 1 generating an emf. which is applied to the input of the picture 
amplifier. 

In designing the mosaic, it is evident that the time-constant of the 
circuit discharging the condenser C must be small enough to allow it 
to discharge fully during the time the beam is on the element. This 
condition requires that C X (R\ + R) be less than the time the beam 
is on the element. In practice, this condition is not difficult to fulfill. 

At this point it is interesting to compare the emf. supplied to the 
amplifier by this storage system with the equivalent voltage from a 



LIGHT 




FIG. 8. Equivalent circuit for non-storage system. 

non-storage system. The equivalent circuit for the non-storage case 
is shown in Fig. 8. The current through the input resistor R 2 will be: 

/.- - 

n 

where F is the light-flux in the picture, 5 the sensitivity of photo- 
sensitive elements, and n the number of picture elements. The volt- 
age to the input of the amplifier is : 



In the storage case the charge accumulated by the element is: 



May, 1937] 



ICONOSCOPES AND KINESCOPES 



483 



where t p is the picture time or l/N for N pictures per second. When 
the beam strikes the element this charge leaves the condenser, result- 
ing in an average current of 

--s- 

where t e is the time the beam is on the element or l/Nn. This current 

is therefore: 

F.sNn 



PHOTOSENSITIVE 




ELECTRON GUM 



FIG. 9. Pictorial representation of conditions on surface of mosaic. 
and the voltage to the amplifier will be : 



Comparing the signal voltages generated in the two cases, we see that 
the ratio is : 

Ki FsRi 
V* ~ FsRi ~ 

n where RI = RI 

Where the number of picture elements is large, as is the case in 
pictures with good definition, this gain in signal is extremely impor- 
tant. For example, the ratio in the case of a 441-line picture is 260,- 
000 times the signal that could be obtained from the non-storage case. 

In order to give a pictorial idea of the conditions on the surface of 
the mosaic, Fig. 9 is included. It represents the appearance of the 



484 



V. K. ZWORYKIN 



[J. S. M. P. E. 



charged image on the mosaic if it were visible to the eye. The region 
just behind the scanning beam is an equilibrium potential and there- 
fore shows no visible image. As we examine the mosaic farther away 
from the line just scanned, we find that the charged image becomes 
more and more intense since the elements have been charging for a 
greater length of time. Just ahead of the scanning beam the image 
reaches its maximum intensity. 

The picture just drawn of the operation of the Iconoscope is very 



FIG. 10. Cathode-ray oscillograph measurements of po- 
tential distribution over mosaic. 

much simplified. A number of factors complicate this seemingly 
straightforward cycle. Among the most important of these compli- 
cating factors are the potential distribution over the mosaic and the 
redistribution of secondary electrons emitted from the elements under 
bombardment. If the average potential of the mosaic is measured in 
darkness while it is being scanned, it will be found to be between 
and 1 volt negative with respect to the electrode which collects the 
electrons leaving the mosaic ; that is, with respect to the second anode. 
However, the potential is not uniform over the surface of the mosaic. 
Elements directly under the beam are found to be in the neighborhood 



May, 1937] ICONOSCOPES AND KINESCOPES 485 

of 3 volts positive with respect to the second anode. As we investi- 
gate elements which have previously been bombarded, we find them 
less positive, until at a point one-quarter to one-third of the vertical 
distance along the mosaic from the point being bombarded, the poten- 
tial has reached l l /z volts negative with respect to the collector. 
The rest of the mosaic is found to be at l 1 /^ volts. Cathode- 
ray oscillograph measurements of the potential distribution over 
the mosaic shows that it can be mapped somewhat as shown in 
Fig. 10. 

In order to account for the potential distribution over the surface 
of the mosaic, it is necessary to consider what takes place among the 
secondary electrons emitted from the cesiated silver elements under 
bombardment. It is well known that when a cesiated silver surface is 
bombarded by an electron beam of the order of 1000 volts velocity, a 
secondary emission of 7 or more times the primary bombarding cur- 
rent can be collected. However, since the mosaic elements are insu- 
lated they must assume, when in equilibrium, a potential such that 
the secondary emission current equals the bombarding current. This 
potential is found to be about 3 volts positive with respect to the 
second anode. In the case of the mosaic in darkness it is obvious that 
the average secondary emission current leaving the mosaic must also 
be equal to the beam current, since the mosaic is an insulator. Thus 
it must come to an equilibrium potential such that the average current 
escaping to the second anode equals the bombarding current. 

Perhaps we should digress at this point and discuss more fully the 
mechanism by which an element acquires this positive equilibrium 
potential. Measurements of the velocity distribution of the sec- 
ondary electrons from a bombarded surface show that they can be 
represented by a distribution curve such as shown in Fig. 11. In the 
figure the abscissa gives the velocity in electron volts of the emitted 
electrons, while the ordinates give the current-per-volt range in ve- 
locity composed of electrons having a given velocity. If the target 
is surrounded with an electrode to collect secondary electrons, the 
current it can collect will depend upon its potential relative to the 
target. When this collector is at zero potential the current reaching 
it will equal the total secondary emission as represented by the total 
area under the curve in Fig. 11. As the collector is made more nega- 
tive, the current decreases, since some of the electrons leaving will 
not have sufficient velocity to reach the electrode and will be driven 
back to the target. The current reaching the collector at some nega- 



486 V. K. ZWORYKIN [J. S. M. P. E 

live potential Vi will be given by the area under the distribution 
curve from Vi to the highest velocity; in other words : 



As the potential of the collector is decreased further eventually it 
will reach a point where the current collected just equals the current 
in the primary beam. At this point no current flows in the external 
lead to the target under bombardment. Experiment shows that for 
a cesiated surface such as is used to make up the globules on the Icono- 
scope mosaic, this potential is in the neighborhood of 3 volts. Hence, 
if an insulated target such as a mosaic element is bombarded, more 




FIG. 1 1 . Velocity distribution of secondary electrons from a bombarded 

surface. 



electrons will leave than will arrive, until the element reaches 3 volts 
positive, at which potential the element will be in equilibrium and the 
currents arriving and leaving will be equal. 

When the mosaic elements are scanned the secondary emission 
from them may be divided into three parts, one going to the second 
anode, another returning to the element itself, and a third being re- 
distributed over the entire mosaic. This latter group which returns 
to the mosaic comes back as a more or less uniform rain of electrons 
having a maximum velocity of about l l / 2 volts. 

This can be verified by removing a portion of the mosaic and sub- 
stituting a metal sheet electrode in its place. If the mosaic, except 
for the substituted portion, is scanned and the current to the metal 
electrode measured, it will be found to decrease as the potential of 



May, 1937] ICONOSCOPES AND KINESCOPES 487 

the probe is decreased. At 1 J /2 volts negative the current will be 
dropped to zero. 

Let us now consider the operation of the Iconoscope in the light of 
the phenomena just discussed. In the first place, due to the poten- 
tial of the mosaic, there is very little electrostatic field aiding the es- 
cape of photoelectrons from the illuminated elements. This means 
that the charging of the globules is dependent in a large measure 
upon the initial velocities of the electrons. Therefore, the photo- 
electric emission is not very efficient, and becomes less so as the illumi- 
nation is increased. It should be remembered, however, that the 
photoemission occurs during the entire picture time, the charge being 
accumulated upon the condensers formed by the silver beads and the 
signal plate. In other systems, since photoemission occurs only 
during the time a picture element is being scanned, there is a gain 
of the order of 10 5 to be had by using the storage system, so that even 
if the above-mentioned photoelectric inefficiency were insurmount- 
able there is still a very great advantage in favor of the Iconoscope 
system. 

As was pointed out in the discussion of the potential distribution, 
there is a line across the mosaic directly behind the scanning beam 
that is 3 volts positive with respect to the second anode, while just 
ahead of the beam the potential is in the neighborhood of l*/2 volts 
negative. There is, therefore, just ahead of the scanning beam, a 
row of elements which have a strong field aiding the leaving photo- 
electrons. This field very much increases the photosensitivity along 
this line and gives rise to a phenomenon known as line-sensitivity. 
This phenomenon can be demonstrated very strikingly in the follow- 
ing way : 

The image from a continuously run motion picture film (i. e. t by 
removing the intermittent and shutter from a motion picture pro- 
jector) is projected upon the mosaic of the Iconoscope. The film is 
run at such a rate that the frame speed is equal to the picture fre- 
quency of the Iconoscope, and in a direction such that the image moves 
oppositely to the vertical direction of scanning. Under these condi- 
tions we find that the Iconoscope transmits a clear image of two 
frames of the motion picture film although, to the eye, there appears 
to be only a blur of light upon the mosaic. 

Thus we have two sources of signal: one, the stored charge over 
the entire mosaic surface; the second, from the sensitive line at the 
scanning beam. At low or normal light intensities, by far the greater 



488 



V. K. ZWORYKIN 



[J. S. M. P. E. 



part of the signal comes from surface sensitivity; but under high 
illumination as much as 50 per cent of the signal may come from line- 
sensitivity. 

As was pointed out above, due to the fact that the secondary emis- 
sion is not saturated, some electrons from the point where the beam 
strikes the mosaic have not sufficient velocity to leave the mosaic 
entirely, but return to its surface as a shower of low-velocity elec- 
trons. The redistributed electrons act to some extent as a high re- 
sistance, short-circuiting the elements, since an element that is more 





Iconoscope Response 

1. Ho background. 

2. 5 inllllluneM/cjn' 
). 10 mllli lumen s/c* 
If. JO miUUumens/cn* 



25 50 75 

Light Intensity (Mi LHlunens/cm*) 

FIG. 12. Response of Iconoscope under various conditions of 
illumination. 

positive than its neighbors tends to receive a greater share of these 
electrons. This resistance is, in effect, identical with that of the dy- 
namic resistance of a triode tube, and under normal operating condi- 
tions is high enough so that it does not produce a very serious loss 
in efficiency; but under high illumination, where considerable differ- 
ence in charge between nearby elements may be developed, this 
shunting resistance may become quite low with the result that there 
is a fairly large loss of signal. 

This redistribution of electrons is, furthermore, responsible for the 
generation of a spurious signal. It appears as an irregular shading 
over the picture, even when the mosaic is not illuminated. The 



May, 1937] 



ICONOSCOPES AND KINESCOPES 



IN!) 



cause of this signal is the variation in instantaneous secondary emis- 
sion current escaping from the mosaic to the second anode. As has 
been pointed out, the average secondary emission from the mosaic 
must be unity, but when we consider that a certain fraction of the 
secondary electrons from the point under bombardment returns to 
other parts of the mosaic, it is quite apparent that the instantaneous 
current leaving the mosaic may vary from point to point. This varia- 
tion is produced by the lack of uniformity of potential and space 
charge over the mosaic. 

It is interesting to note that if a clean sheet of metal is substituted 
for the mosaic, the spurious signal appears when it is scanned, pro- 
vided the secondary emission is not saturated. The signal disap- 




FIG. 13. Diagram of typical Kinescope. 

pears, however, if the metal plate is made sufficiently positive or 
negative with respect to the second anode. Under these conditions, 
the secondary emission is either saturated or suppressed. In prac- 
tice the effect of this signal can be eliminated by the introduction of a 
compensating signal. The spurious signal varies rapidly with beam 
current, and under conditions of low beam intensity and moderately 
high illumination it is negligible compared with the picture signal. 

So far, the scanning beam has been considered as some sort of 
commutating switch which sweeps over the mosaic. Actually, the 
beam does behave in just this way. The beam, when falling upon an 
element, connects it through a resistance (dynamic, of course) to the 
second anode. This is obvious when we consider the action of the 
beam. As has been pointed out, the ratio of secondary electrons to 
primary electrons from a cesiated silver surface is about 7 when 



490 



V. K. ZWORYKIN 



[J. S. M. P. E. 



saturated. However, if the bombarded surface is made positive this 
ratio decreases, reaching unity at +3 volts and one-half at about 10 
volts. From curves giving the secondary emission ratio of an element 
for various collector potentials, together with a knowledge of the beam 
current, the effective resistance connecting the bombarded element 
with the second anode can easily be estimated. This resistance 




-o 3 
o a 2 



.Volts 



FIG. 14. Typical control characteristic of 
Kinescope gun. 

turns out to be of the order of 10 6 ohms. If the beam current is too 
weak, it will not fully restore the illuminated element to equilibrium. 
Considering a stationary picture, and neglecting the effect of the redis- 
tribution of scattered electrons, this would not reduce the signal ob- 
tained from a given amount of light. However, it would cause a 
lag and consequent blurring of the image of a moving object. In the 
actual Iconoscope, because of the role the beam plays in establishing 
the potential of the mosaic, and because of the shunting effect of re- 



May r 1937] ICONOSCOPES AND KINESCOPES 491 

distributed electrons, there is an optimal beam current at which the 
signal is a maximum for a given condition of light. 

Taking into account the various factors tending to reduce the out- 
put of the Iconoscope, it is found that the net efficiency of conversion 
is in the neighborhood of 5 to 10 per cent. In other words, the signal 
output is about Vao that which would be expected upon the basis of 
the light-flux reaching the mosaic, the saturated photoemission of 
photoelectric elements, and the assumption that the entire photo 
current is stored by the mosaic. The efficiency of conversion is not 
constant, but as explained above depends upon the amount of light 
used. The efficiency is a maximum at low light and decreases as the 



10 corresponds 4o fluorescence Juri 




20 120 (,0 1Q 30 20 

Time in Seconds 

FIG. 15. Phosphorescent decay curve for zinc orthosilicate. 

light is increased. This point will be considered again under a dis- 
cussion of the actual performance of the Iconoscope. 

Up to this point we have based our consideration of the relative 
merits of the storage and non-storage types of systems upon a com- 
parison of signal output alone. The recent development of the sec- 
ondary emission multiplier makes it necessary to introduce other 
considerations into this comparison. The electron multiplier pro- 
vides a means of amplifying a photoelectric current to almost any 
desired extent without introducing any additional "noise" into the 
signal. It might seem, therefore, that we could amplify the minute 
photo current obtained by the conventional scanning system to such 
an extent that the sensitivity of the two systems would be equal. 
This, however, can not be done, because even with a perfect amplify- 



492 V. K. ZWORYKIN [J. S. M. P. E. 

ing system the statistical fluctuations in the original photo current are 
amplified just as much as the signal is amplified. Because of this 
there is a definite limit to the sensitivity of this type of system im- 
posed by the original shot noise in the photo current. In the case of 
the Iconoscope there is a similar limit, but because the charge repre- 
senting each picture element is so much greater than in the non- 
storage case the ideal sensitivity is very much greater. Actually, in 
the type of Iconoscope described in this paper, the limit of the sen- 
sitivity is not set by the statistical fluctuations of the stored charge, 
but by the thermal noise in the coupling resistor to the amplifier. A 
quantitative comparison of the limiting sensitivity of the Iconoscope 




FIG. 16. Kinescope with 9-inch viewing screen. 

used at present, taking into account its inefficiency and imperfections, 
shows that it is able to operate at one-tenth the light required by a 
perfect non-storage system. This, of course, includes the use of an 
electron multiplier, and applies to electrically scanned as well as 
mechanically scanned non-storage systems. 

The resolution of an Iconoscope may be limited either by the size 
of the photoelectric elements or by the size of the scanning beam. 
The size of the silver globules in the Iconoscope described is many 
times smaller than a picture element, so that many hundreds of them 
act together under the scanning spot. The resolution is limited, 
therefore, by the spot size. At present, the resolution adopted is 
about 441 lines, but when necessary the beam size can be reduced and 
the resolution made much higher. 



May, 1937] 



ICONOSCOPES AND KINESCOPES 



493 




FIG. 17. Televised picture upon the screen of the Kinescope. 
FIG. 18. Another televised picture. 



494 



V. K. ZWORYKIN 



[J. S. M. P. E. 



The actual response of an Iconoscope under various conditions of 
illumination is shown in Fig. 12. The output is measured in milli- 
volts across a 10,000-ohm coupling resistance and the light input 
measured in lumens per square centimeter on the mosaic. The curve 
showing the greatest response represents the signal output from a 
small illuminated area when the remainder of the mosaic is in dark- 
ness. The other curves of the family show the response from the 
same area when the mosaic is illuminated with a uniform background 
of light. The response is not linear, but falls off as the illumination 
is increased until it reaches a saturation value. The saturation volt- 
age output is nearly constant for tubes of a given design, but the 
slope of the response curve may vary from tube to tube, depending 
upon treatment, and is a measure of the sensitivity. 

The decrease in sensitivity with illumination is not wholly disad- 
vantageous in that it permits the transmission of a wider range of 
contrast over a given electrical system than would otherwise be pos- 
sible. In a sense, this is similar to the compressor-expander systems 
used in sound recording. 

In spite of the complicated manner of its operation and the factors 
mentioned reducing its efficiency, the Iconoscope is an extremely sen- 
sitive and stable device for television transmission. Excellent and 
consistent results are obtained under widely varying conditions of 
operation. The practical lower limit to light that can be used to 
transmit a picture is set by the "noise" in the picture amplifier. 
Measurements have been made to determine the illumination neces- 
sary for satisfactory operation. With an//2.7 lens to focus the image 
upon the mosaic, an average surface brilliancy of from 30 to 50 
candles/sq. ft. on the object viewed gives completely satisfactory 
transmission. A recognizable image can be obtained from a good 
Iconoscope with 8 candles/sq. ft. using an //16 lens; that is, with 
VIM the illumination mentioned above. 

For comparison, the illumination of some scenes commonly met 
with is given in the following table : 



Scene 


Location 


Date 


Time 


Weather 


Brightness 


Beach 


Atlantic City 


August 


2:00 P.M. 


Hazy 


500 


Boardwalk 


Atlantic City 


August 


2:00 P.M. 


Hazy 


275 


Street 


Philadelphia 


August 


2:30 P.M. 


Clear 


200 


Times Square 


New York City 


November 


1:30 P.M. 


Rain 


40 


Street Parade 


East Orange 


November 


10:30 A.M. 


Rain 


40 to 60 



May, 1937] 



ICONOSCOPES AND KINESCOPES 



It is evident that perfectly satisfactory outdoor pick-up may be 
obtained under almost all average conditions of light. 

The device used to reproduce the television picture is also electron 
operated. This tube, which has been named the "Kinescope," is 
similar to a cathode-ray oscilloscope in many respects. It consists of 
an electron gun for defining and 
controlling a cathode-ray beam, 
and a fluorescent screen which 
becomes luminous under bom- 
bardment from the electron gun. 
A diagram of a typical Kinescope 
is shown in Fig. 13. 

The cathode-ray beam is made 
to sweep across the fluorescent 
screen in synchronism with the 
scanning beam in the Iconoscope 
which is transmitting the picture. 
Furthermore, the current in the 
Kinescope cathode-ray beam is 
controlled by the signal impulses 
generated at the Iconoscope. 
This control acts in such a way 
that the impulse corresponding 
to a bright area on the Icono- 
scope causes an increase in cur- 
rent, while a dark region causes 
a decrease. There will, there- 
fore, be an exact correspondence 
both in position and intensity 
between the fluorescent illumina- 
tion on the Kinescope screen 
and the light on the mosaic in 
the Iconoscope. A picture projected on the Iconoscope will therefore 
be reproduced by the Kinescope. 

The electron gun in the receiving tube is similar in principle to that 
in the Iconoscope, but is made to handle larger currents and to oper- 
ate at higher voltages. Furthermore, since the picture is repro- 
duced by modulating the beam current, the control grid is a much 
more critical item. The control-grid characteristic is determined by 
a number of factors, such as the grid aperture, the spacing and geom- 




FIG. 19. A typical studio pick-up 
camera employing the Iconoscope. 



496 



V. K. ZWORYKIN 



[J. S. M. P. E. 



etry of the cathode, the first anode, etc. Fig. 14 shows a typical con- 
trol characteristic for a Kinescope gun. 

The fluorescent screen is made by coating the flat portion of the 
glass bulb with a synthetic zinc orthosilicate, very similar to natural 
willemite. The synthetic material has high luminous efficiency, the 

light output at a given voltage 
being proportional to the cur- 
rent striking it. At 6000 volts 
the material gives nearly 3 
candles per watt. The effi- 
ciency of light production 
varies somewhat with voltage 
used, but at higher beam 
velocities is nearly constant. 
This can be seen from the 
general relation between 
candlepower P, current inten- 
sity /, and applied voltage V, 
which is given by the equation : 

P = AI( V - Fo) 

A is a constant depending 
upon the phosphorescence and 
V the extrapolated minimum 
exciting voltage, which proves 
to be in the neighborhood of 
1000 volts. 

In addition to its high lumi- 
nous efficiency, this material 
does not burn or disintegrate 
under bombardment with elec- 
trons. The phosphorescent 

properties of a fluorescent material are an important consideration. 
An ideal substance for television work should emit a constant 
amount of light for one entire picture frame and drop to zero at 
the end of this period. If the phosphorescence time is too long, the 
moving portions of a picture will leave a ' 'trail. ' ' For example, the path 
>f a moving ball will be marked with a comet-like tail. On the other 
hand, if the decay time is too short, flicker becomes noticeable. The 
phosphorescence decay curve for zinc orthosilicate is shown in Fig. 15. 




FIG. 20. 



Console type of 
receiver. 



television 



May, 1937] ICONOSCOPES AND KINESCOPES 497 

Fig. 16 shows a photograph of a Kinescope with a 9-inch view- 
ing screen. This is only one of a number of sizes, both larger and 
smaller, and designs possible for the Kinescope. 

Between the transmitting Iconoscope and the reproducing Kine- 
scope there is a chain of electrical equipment involving the picture 
amplifier, transmitter, radio receiver, and synchronizing system. 
This field is much too large to cover in this paper and has been treated 
in detail elsewhere. 2 

In closing, it might be well to illustrate the performance of the 
systems with some photographs of televised pictures as they appear 
upon the screen of the Kinescope. These are shown in Figs. 17 and 
18. The appearance of a typical studio pick-up camera using the 
Iconoscope is shown in Fig. 19, while that of a console type television 
receiver can be seen in Fig. 20. 

REFERENCES 

1 ZWORYKIN, V. K.: U. S. Patent No. 1,691,324: "Television System." 
1 ZWORYKIN, V. K. : "An Experimental Television System and Kinescope," 
Proc. I. R. ., 21 (Dec., 1933), No. 12, p. 1655. 

Symposium on an Experimental Television System: 

KELL, R. D., BEDFORD, A. V., AND TRAINER, M. A.: "Transmitter," p. 1246. 
HOLMES, R. S., CARLSON, W. L., AND TOLSON, W. A.: "Receivers," p. 1266. 
YOUNG, C. J.: "Radio Link for Television Signals," Proc. I.E.R., 22 
(Nov., 1934), No. 11, p. 1268. 



NEW MOTION PICTURE APPARATUS 



During the Conventions of the Society, symposiums on new motion picture appara- 
tus are held, in which various manufacturers of equipment describe and demonstrate 
their new products and developments. Some of this equipment is described in the 
following pages; the remainder will be published in subsequent issues of the Journal. 



A SINGLE-CHANNEL RECORDING AND RE-RECORDING 
SYSTEM* 



H. I. REISKIND** 

When sound motion pictures were first introduced, quite a large number of 
recording installations were made in the East. Since that time, however, very 
few completely new recording channels have been installed. Although the old in- 
stallations have been modernized, it has not always been practicable to include 
all the operating conveniences that the advances in the recording art indicated as 
being desirable. For that reason a description of the new installation of RCA 
recording equipment in the Twentieth Century-Fox Film studio at New York 
may be in order. 

In this, as in all eastern studios, the requirement that the space occupied be 
kept at a minimum for the facilities furnished, was of extreme importance; and 
at the same time it was, of course, necessary that the recording and re-recording 
facilities be on a par with those in the studios on the West Coast. 

In discussing this installation an attempt will be made to show the facilities 
furnished, the schematic equipment layouts, and the general installation methods. 
Since most of the units used in this installation have already been described in the 
JOURNAL, 1 this paper will be concerned mainly with the general overall arrange- 
ment of the system. The installation consists of a single channel which pro- 
vides facilities for recording on any one of four stages, with set mixing and head- 
phone monitoring. It is also possible to re-record from four sound-tracks and 
one or more microphones, using a permanent eight-position mixer installed in 
the control room in which complete compensation facilities are provided. A 
complete theater reproducing system is provided for monitoring the re-record- 
ing, which may be used also for stage recording when desirable. Smaller com- 
pound baffles are provided in the recording room and above there-recording 
console. 



* Received February 25, 1937; presented at the Meeting of the Atlantic 
Coast Section, February 25, 1937. 

** RCA Manufacturing Co., New York, N. Y. 
498 



NEW MOTION PICTURE APPARATUS 



499 



Fig. 1 shows the general layout of the studios. The recording and control 
rooms shown were originally part of stage D. The wall between this stage and 
the control room carries a large double-glass window, 3'/i by 7 feet, that can be 
opened, converting the control room and stage D into a monitoring room (Fig. 2). 
When it is necessary to record on stage D, the double windows may be dosed 
and the compound baffle used for monitoring. 

The projection machine located upon the level above the control room projects 
a picture upon a screen above the projection speakers. The projector may also 
be turned about so as to project a picture upon a screen on stage A for orchestra 
scoring 01 similar work. This projector is fitted with a preview attachment that 



Stage A. 



Projection 




Stage D 



Stage C 



FIG. 1. General layout of the studio, showing control room, recording 
room, projection room, and three stages. 



makes it possible to run separate picture and sound-track on the same machine. 
Together with the two theater-type sound-heads and the recorder-type film 
phonograph, a picture can be shown and sound reproduced from four sound- 
tracks. 

The recording and projection amplifier racks are set into the wall, thus saving 
considerable floor space. This was made possible by the front-servicing con- 
struction of these amplifiers. All component parts are mounted upon the front 
panel, which may be opened on hinges without disturbing any of the wiring. 
Locating the two amplifiers at the sides of the console brings all the adjustments 
within reach of the monitor man. The telephone, signal lights, and motor- 
starting positions are on the right of the recording machine, within easy reach 
of the operator. 



500 



NEW MOTION PICTURE APPARATUS [J. S. M. P. E. 



Referring to Fig. 3, the outputs of the four microphone and four phototube 
pre-amplifiers are fed into an eight-position mixer and compensator. The output 
of the re-recording console, which will be examined in detail later, is fed to a low- 
pass filter adjustable to nominal cut-off values of 6000, 8000, or 10,000 cps. The 
output of the filter is fed to the recording amplifier, the output of which, in turn, 




FIG. 2. View of re-recording console, showing double 
glass window, which may be opened into stage D. At the 
top may be seen the compound baffle used for monitoring, 
and the window through which pictures may be projected 
upon the screen in stage D from the projection room (on the 
upper floor at the right, not shown in the picture). 



operates the recording galvanometer. Since the internal impedance of the gal- 
vanometer varies with frequency, the noise-reduction amplifier and monitor- 
ing systems derive their input signal currents from a tertiary winding on the 
input transformer feeding the output stage of the recording amplifier. The neon 
volume indicator is fed from a separate output on the recording amplifier through 
a resistive network incorporated in the recording amplifier. The monitoring 
amplifier is a-c. operated, and supplies signals to the high-quality head-phones 



May, 1937] 



NEW MOTION PICTURE APPARATUS 



501 



and to the compound baffles through a distribution network. The adjustment 
of the noise-reduction amplifier, described previously in the JOURNAL, is facilitated 
by push-buttons upon the front panel. 

The monitoring decompensator consists of a resistive network and a single- 
section filter. The filter introduces a high-frequency loss that closely approxi- 
mates film-transfer and reproducer-slit loss. By using a standard theater pro- 
jection amplifier and high-quality theater speakers, the monitoring quality closely 
approximates the quality of the final print as reproduced in the theater. 

The recording machine employs the well known magnetic drive and an optical 
system having interchangeable intermediate lens barrels. By using these dif- 




FIG. 3. Set-up of channel for re-recording. 

ferent barrels, it is possible to record any of a number of types of track, such as 
symmetrical variable-width track with bias noise reduction, split-wave (class B 
push-pull), track with shutter noise reduction, and class A push-pull. All these 
types of track are recorded with ultraviolet light, resulting in tracks of much 
better high-frequency definition. Should it be desirable for any reason, variable- 
density tracks may also be recorded, using still another intermediate lens 
assembly. 

The film phonograph indicated in Fig. 3 employs the same magnetic drive as 
the film recording machine. All the sound-heads are of the standard theater 
type, using rotary stabilizers; and all the machines will reproduce either standard 
or push-pull track. 

Fig. 4 is a diagram of the re-recording console, which consists of the necessary 
microphone and phototube pre-amplifiers, two four-position mixers with asso- 



502 



NEW MOTION PICTURE APPARATUS [J. S. M. P. E. 



ciated compensators, two booster amplifiers, and a combining network. The 
outputs of the mixers are amplified by the booster amplifiers and combined in the 
resistance-transformer network. The mixing potentiometers used in both this 
mixer and in the "tea-wagon" console are bridged- T networks, which have the ad- 
vantage of maintaining constant impedance regardless of the attenuation setting. 
The four inputs are coupled to the mixing potentiometers through transformers, 
and the output is coupled to the transmission line in the same way. This pro- 
vides for balancing the transmission lines to ground, and reduces the possibility 
of pick-up and cross-talk. Since the overall master control is seldom used in 
re-recording, the panels have been laid out with the four mixer knobs in a line at 



o Sound 
Heads 



V" 

V 


2-Pos. 

V^r. Att. 


"V 

v 




P ^ 






TV vy 


4-Pos. 




T^ ^T 


Mixer 




jv vy 







FIG. 4. Transmission diagram of re-recording console. 

the bottom, and the master at the top center of the panel where it will be out 
of the way. Off -on key switches are provided for each mixer position, at the top 
of the panel on either side of the master control. These switches are of the toggle 
type and have three positions: a central off position, and two on positions. 
These two on positions differ in that the key locks in place in the forward position, 
but in the back position it must be held. The back position may be used instead 
of a push-button control for certain effects such as water splashes, explosions, etc. 
The units shown as additional attenuators consist of two balanced-tf variable 
attenuators, terminating in jacks which may be patched into any input circuit. 
They are controlled by key switches identical to those in the re-recording mixer. 
These attenuators are often useful for obtaining very large level ranges for certain 
effects, and for providing given amounts of attenuation which may be pre-set and 
inserted upon cues. 



May, 1937] 



NEW MOTION PICTURE APPARATUS 



503 



The four-position compensator shown in Fig. 4 provides the mixer man with 
complete control of the frequency characteristic of each track. In the majority 
of cases the frequency characteristic should be set by means of formal tests and 
then left alone; but many cases occur in re-recording in which it is desirable to 
change the frequency characteristic to correct for poor recording on particular 




Fig. 5. 




/ 

W 



. 7 




FIG. 5. Characteristic of low-frequency section of four- 
position compensator. 

FIG. 6. Characteristic of high-frequency section of com- 
pensator. 

FIG. 7. Characteristic of middle-range section of com- 
pensator. 



takes. The compensator is unique in that the frequency range has been divided 
into three bands, a separate section of the compensator operating on each of 
the bands. 

Fig. 5 shows the characteristic of the low-frequency section in the positions 
of maximum compensation. There are two additional steps of "raise low fre- 
quencies" and "reduce low frequencies" that are less drastic than the curves 
shown. The six low-frequency characteristics and a position of zero compensa- 
tion are controlled by a dial switch. 



504 



NEW MOTION PICTURE APPARATUS [J. S. M. P. E. 



Fig. 6 shows the characteristic of the high-frequency section of the compen- 
sator. Here again three conditions of "raise" and three conditions of "reduce 
high frequencies" are provided. 

Fig. 7 shows the characteristic of the middle-range section, which operates 
over the frequency band between 200 and 3000 cps. and has proved extremely 
useful for salvaging very bad recordings. The characteristics of the three sec- 
tions are all additive, and almost any desired overall characteristic may be at- 
tained. A key switch is provided so that the compensator may be inserted into 
or removed from the circuit. When removed, it is replaced by a fixed pad equal 




-d 
-m 



Proj. 
Amp. 


- 


Proj. 
Sp'kr 
Syst. 



FIG. 8. Switching diagram. 

in loss to the insertion loss of the compensator, and thus constant level into the 
mixer is maintained at all times. 

The various units shown in Fig. 4, i. e., mixers, compensators, microphone and 
phototube pre-amplifier, etc., are all located in the re-recording console. The 
three sets of amplifiers are mounted in the console on shock-proof shelves, and 
may be readily removed for servicing since connections to them are made through 
Cannon plugs and receptacles. 

The input to the microphone pre-amplifier comes directly from the stage dis- 
tribution box, one of which is located on each stage. This furnishes a mounting 
for three rows of Cannon receptacles. The lines from the microphone pre-ampli- 
fiers terminate in one of these rows of receptacles. When it is desired to record 
a microphone pick-up through the re-recording console, the microphone may be 
plugged into the proper receptacle in the stage distribution box and the signal 
thus fed directly to the pre-amplifier. The various pre-amplifier outputs and the 
mixer inputs appear upon a jack bay which is part of the re-recording console and 
which allows the selection of any desired mixer. If, because of severe fields upon 



May, 1937] NEW MOTION PICTURE APPARATUS 505 

the stage, it is desired to locate the microphone prc-amplifier nearer the micro- 
phone, the pre-amplifier may be removed from the console and plugged into one 
of the other sets of receptacles on the stage distribution panel. The output of 
the pre-amplifier then appears upon one of the central sections of jacks and is 
patched to the mixer. The B batteries supplying plate voltage to the microphone, 
phototube, and booster amplifiers are all located in shielded compartments in the 
re-recording console. 

When mixing on the set is desired, the portable "tea-wagon" console is used. 
The microphone pre-amplifiers are removed from the re-recording console and 
inserted into the "tea-wagon" console. This portable console is connected on 
any stage by using the bottom set of receptacles on the stage distribution panel. 
The overall circuit arrangement is identical to that used in re-recording, except 
that, due to the low insertion loss of the portable mixer and compensator, the 
booster amplifiers are unnecessary. The console provides a mounting for the four 
microphone pre-amplifiers, volume indicator, and a four-position mixer. Each 
mixer position has individual high- and low-frequency compensation. Normally 
the low-frequency compensation is either 9 or 13 db. at 100 cps., and is selected 
by a three-position key switch. By a change in wiring these values may be 
changed to 5 and 9 db. The high-frequency compensation begins at approxi- 
mately 5000 cps. and increases to 10 db. at 10,000 cps. 

In order to simplify operation and to change recording equipment set-ups as 
quickly as possible, it was decided to eliminate jacks and patch cords wherever 
possible. 

Fig. 8 shows, in extremely simplified form, a schematic diagram of the equip- 
ment and the switching arrangements. The three switches A 1 , A 2, and A 3 are, in 
reality, a single three-position switch mounted at the top of the recording ampli- 
fier rack. Switch C is mounted in the projection amplifier rack, and switch B 
upon the projection machine. In the positions shown in Fig. 8 the channel is set 
up for re-recording. The output of the re-recording console is connected to the 
recording amplifier rack, and the projection system is used for monitoring. 

For stage recording, switch Al is moved to the up position, leaving all the 
others as shown. This provides the same set-up as in the previous case, merely 
substituting the portable "tea-